Beruflich Dokumente
Kultur Dokumente
FASTFIND LINKS Document Organization Product Version Getting Help Table of Contents
FE-98ARC018-07
Copyright 20082011 Hitachi Data Systems Corporation, ALL RIGHTS RESERVED No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or stored in a database or retrieval system for any purpose without the express written permission of Hitachi Data Systems Corporation (hereinafter referred to as Hitachi Data Systems). Hitachi Data Systems reserves the right to make changes to this document at any time without notice and assumes no responsibility for its use. This document contains the most current information available at the time of publication. When new and/or revised information becomes available, this entire document will be updated and distributed to all registered users. Some of the features described in this document may not be currently available. Refer to the most recent product announcement or contact your local Hitachi Data Systems sales office for information about feature and product availability. Notice: Hitachi Data Systems products and services can be ordered only under the terms and conditions of the applicable Hitachi Data Systems agreement(s). The use of Hitachi Data Systems products is governed by the terms of your agreement(s) with Hitachi Data Systems. By using this software, you agree that you are responsible for: a) Acquiring the relevant consents as may be required under local privacy laws or otherwise from employees and other individuals to access relevant data; and b) Ensuring that data continues to be held, retrieved, deleted, or otherwise processed in accordance with relevant laws. Hitachi is a registered trademark of Hitachi, Ltd. in the United States and other countries. Hitachi Data Systems is a registered trademark and service mark of Hitachi, Ltd. in the United States and other countries. All other trademarks, service marks, and company names are properties of their respective owners.
Contents
Preface........................................................................................................ xi
Intended audience . . . . Product version . . . . . . Document organization . Syntax notation . . . . . . Related documents. . . . Getting help. . . . . . . . . Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... ........... ........... ........... ........... ........... ........... . . . . . . . . . . . . . . . . . . . . . .... .... .... .... .... .... .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi . .xi . xii . xiii . xiv xvii xvii
HCP overview.....................................................................................1-1
Introduction to Hitachi Content Platform. . . . . . . . . . . . . . . HCP hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nodes and storage . . . . . . . . . . . . . . . . . . . . . . . . . . . Storage nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . Search nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Networking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . System architecture . . . . . . . . . . . . . . . . . . . . . . . . . . HCP features that depend on hardware configuration. . . HCP software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HCP System Management Console . . . . . . . . . . . . . . . . . . . HCP service activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 . 1-4 . 1-5 . 1-5 . 1-5 . 1-6 . 1-6 . 1-9 .1-10 .1-11 .1-11
iii
. . . .
. . . .
. . . .
. . . .
. . . .
.. .. .. ..
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
iv
Online and offline upgrades. . . . . . . . . . . . . . . . . . . . . . . . Considerations for online upgrades. . . . . . . . . . . . . . . . Monitoring HCP during an online upgrade . . . . . . . . . . . Before you begin an HCP upgrade . . . . . . . . . . . . . . . . . . . Information you need for an upgrade . . . . . . . . . . . . . . Resources you need for an upgrade . . . . . . . . . . . . . . . Running upgrade prechecks . . . . . . . . . . . . . . . . . . . . Freeing space on the HCP nodes . . . . . . . . . . . . . . . . . Preparing the HCP system for an upgrade. . . . . . . . . . . . . . Performing the upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . Step 1: Log in for the upgrade . . . . . . . . . . . . . . . . . . Step 2: Retrieve the files for the upgrade. . . . . . . . . . . Step 3: Set up the upgrade . . . . . . . . . . . . . . . . . . . . Step 4: Execute the upgrade . . . . . . . . . . . . . . . . . . . Step 5: Verify the upgrade . . . . . . . . . . . . . . . . . . . . . Step 6: Complete the upgrade . . . . . . . . . . . . . . . . . . Step 7 (conditional): Resume replication . . . . . . . . . . . Restarting an upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . Rolling back an offline upgrade . . . . . . . . . . . . . . . . . . . . . Resources you need for an upgrade rollback . . . . . . . . . Performing the upgrade rollback . . . . . . . . . . . . . . . . . Step 1: Log in for the upgrade rollback . . . . . . . . . . Step 2: Execute the upgrade rollback . . . . . . . . . . . Step 3: Verify the upgrade rollback . . . . . . . . . . . . . Installing a hotfix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resources you need for a hotfix installation . . . . . . . . . Performing the hotfix installation . . . . . . . . . . . . . . . . . Step 1: Log in for the hotfix installation . . . . . . . . . . Step 2: Execute the hotfix installation . . . . . . . . . . . Step 3: Verify the hotfix installation. . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 4-3 . 4-3 . 4-5 . 4-5 . 4-6 . 4-6 . 4-7 .4-14 .4-16 .4-17 .4-17 .4-19 .4-22 .4-23 .4-24 .4-25 .4-27 .4-27 .4-29 .4-30 .4-30 .4-30 .4-31 .4-32 .4-32 .4-33 .4-33 .4-34 .4-34 .4-36
Step 4: Set up the node addition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-11 Step 5: Execute the node addition. . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-12 Verifying the node addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-13
vi
Verifying the storage addition . . . . . . . . . . . . . . . . . . . . . . . . . Reconfiguring HCP to use a new or different storage array . . . . . . . . Using TrueCopy to migrate data. . . . . . . . . . . . . . . . . . . . . . . . Step 1: Prepare the new array . . . . . . . . . . . . . . . . . . . . . . Step 2: Use TrueCopy to migrate the data . . . . . . . . . . . . . . Step 3: Reconfigure the SAN fabrics . . . . . . . . . . . . . . . . . . Step 4: Update the HCP nodes to recognize the new array . . Step 5: Power on the nodes . . . . . . . . . . . . . . . . . . . . . . . . Step 6: Reenable zero-copy failover . . . . . . . . . . . . . . . . . . Virtualizing an existing storage array behind an enterprise array . Step 1: Prepare the storage arrays for the virtualization . . . . Step 2: Prepare the HCP system for the virtualization . . . . . . Using Tiered Storage Manager to migrate data . . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
.7-10 .7-11 .7-12 .7-13 .7-13 .7-13 .7-14 .7-15 .7-15 .7-15 .7-16 .7-17 .7-18
vii
Enabling or disabling support for replication . . . . . . . . . . . . . . . . . . . . . . . . .9-16 Enabling support for replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-16 Disabling support for replication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-17
viii
C D
HCP configuration file........................................................................ C-1 Setting up SSH key authentication .................................................... D-1
Adding and forwarding an SSH key in Windows. . . . . . . . . . . . . . . . . . . . . . . D-2 Adding and forwarding an SSH key in Linux . . . . . . . . . . . . . . . . . . . . . . . . . D-3
Glossary Index
ix
Preface
This book is the service guide for Hitachi Content Platform (HCP). It contains complete information for installing and upgrading HCP system software. It also provides instructions for system reconfiguration, management, and troubleshooting procedures not available to customers. This book focuses on the HCP software. For information on maintaining the hardware for an HCP system, see the applicable HCP hardware maintenance documentation. Note: Throughout this book, the word Unix is used to represent all UNIXlike operating systems (such as UNIX itself or Linux), except where Linux is specifically required.
Intended audience
This book is for Hitachi Data Systems (HDS) service and support personnel and for HDS partners and authorized resellers who are responsible for installing and upgrading the HCP software and providing the first tier of customer support. It assumes you are familiar with basic HCP concepts and administrative procedures. It also assumes you have at least a couple of years of experience working with computer systems, including using command-line interfaces.
Product version
This book applies to release 4.1 of HCP.
xi
Document organization
Document organization
This book contains eleven chapters and five appendixes.
Chapter/Appendix
Chapter 1, HCP overview
Description
Contains an overview of HCP hardware and software, as well as an outline of the HCP service provider responsibilities described in this book Describes the information you need to perform a fresh installation of HCP Contains complete instructions for installing an HCP system Contains complete instructions for upgrading an HCP system Contains instructions for the software portion of the procedure for adding nodes to an HCP system Contains instructions for recovering nodes with corrupted file systems Contains instructions for the software portion of the procedure for adding storage to an HCP SAIN system Provides instructions for retiring nodes and storage arrays from HCP systems Provides instructions for various system reconfiguration procedures, most of which you perform through the HCP System Management Console Provides instructions for various system management procedures, most of which you perform through the HCP System Management Console Explains procedures that can help identify or repair problems with an HCP system Contains forms you can use to record the information you gather for an OS or HCP software installation, system upgrade, or node addition; also contains a form for recording the encryption key displayed during a core software installation Shows the menus you use when installing, upgrading, or adding nodes to an HCP system
Chapter 2, Gathering installation information Chapter 3, Installing an HCP system Chapter 4, Upgrading an HCP system Chapter 5, Adding nodes to an HCP system Chapter 6, Maintaining HCP nodes Chapter 7, Managing storage in an HCP SAIN system Chapter 8, Retiring devices Chapter 9, Reconfiguring HCP system software
xii
Syntax notation
(Continued)
Chapter/Appendix
Appendix C, HCP configuration file
Description
Describes all the configuration parameters used in the installation or upgrade of an HCP system Explains how to add SSH keys to a keychain Explains how to access the HCP Configuration menu, the HCP System and Tenant Management Consoles, and the Namespace Browser through the backend network
Appendix D, Setting up SSH key authentication Appendix E, Accessing HCP interfaces through the back-end network
Syntax notation
The table below describes the conventions used for the syntax of commands, expressions, URLs, and object names in this book.
Notation boldface Meaning
Type exactly as it appears in the syntax (if the context is case insensitive, you can vary the case of the letters you type) Replace with a value of the indicated type Vertical bar Choose one of the elements on either side of the bar, but not both Square brackets Include none, one, or more of the elements between the brackets Parentheses Include exactly one of the elements between the parentheses Replace with the combination of the directory path and name of a file
Example
This book shows: --version You enter: --version
italics |
This book shows: node-ip-address You enter: 192.168.210.16 This book shows: -h|--help You enter: -h or: --help This book shows: [--unique-ify] You enter: --unique-ify or nothing This book shows: (-u|--user) username You enter: -u root or: --user root This book shows: source-file-spec You enter: /opt/arc/build.version
[ ]
( )
-file-spec
xiii
Related documents
(Continued)
Notation -path
Meaning
Replace with a directory path with no file or object name
Example
This book shows: destination-path You enter: /opt/arc/tools/cluster/temp
Related documents
The following documents contain additional information about Hitachi Content Platform:
xiv
Related documents
Using the Default Namespace This book describes the file system
HCP uses to present the contents of the default namespace. It provides instructions for accessing the namespace by using the HCP-supported protocols for the purpose of storing, retrieving, and deleting objects, as well as changing object metadata such as retention and permissions.
Using HCP Data Migrator This book contains the information you
need to install and use the HCP Data Migrator (HCP-DM) utility distributed with HCP. This utility enables you to copy data between local file systems, HCP namespaces, and earlier HCAP archives. It also supports bulk delete operations. The book describes both the interactive window-based interface and the set of command-line tools included in HCP-DM.
Using the HCP Client Tools This book contains the information you
need to install and use the legacy set of client command-line tools distributed with HCP. These tools enable you to find files and to copy and move files to and from namespaces. The book contains many examples that show command-line details and the overall workflow. Note: For most purposes, the HCP client tools have been superseded by HCP Data Migrator. However, they have some features, such as finding files, that are not available in HCP-DM.
xv
Related documents
xvi
Getting help
Getting help
The Hitachi Data Systems customer support staff is available 24 hours a day, seven days a week. If you need technical support, please call:
United States: (800) 446-0744 Outside the United States: (858) 547-4526
Note: If the customer purchased HCP from a third party, please contact the applicable HCP support center.
Comments
Please send us your comments on this document: hcp.documentation.feedback@hds.com Include the document title, number, and revision, and refer to specific sections and paragraphs whenever possible. Thank you! (All comments become the property of Hitachi Data Systems.)
xvii
Comments
xviii
1
HCP overview
Hitachi Content Platform (HCP) is the distributed, fixed-content, data storage system from Hitachi Data Systems. An HCP system consists of both hardware and software. Once the HCP hardware is set up, the software must be installed, sometimes upgraded or reconfigured, and always managed. The customer is responsible for some management activities. As a service provider, you are responsible for installing, upgrading, and reconfiguring the software, as well as for system management activities that are unavailable to customers. You are also responsible for adding, maintaining, and removing nodes and storage arrays. This chapter reviews some basic HCP concepts and describes the HCP hardware architecture. It also contains an overview of your responsibilities as a service provider.
11
12
Namespace access protocols HCP provides access to namespaces through a variety of industry-standard protocols. For actions such as adding objects to a namespace, viewing and retrieving objects, changing object metadata, and deleting objects, HCP supports the HTTP, WebDAV, CIFS, and NFS protocols. The latter three protocols are supported only for the default namespace. HCP allows special-purpose access to the default namespace through two additional protocols: SMTP (for storing email) and NDMP (for backing up and restoring data). Data access accounts To access the data in an HCP namespace, users and applications must present valid credentials. These credentials are defined by a data access account. The account specifies a username and password. It also specifies which namespaces the user or application can access and which operations the user or application can perform in each of those namespaces. HCP Search Console HCP includes a web application called the Search Console that lets users search for objects based on specified criteria. This Console works with either of two implementations: The HDDS search facility This facility interacts with Hitachi Data Discovery Suite (HDDS), which performs searches and returns results to the HCP Search Console. To use the HDDS search facility, you need to first install and configure HDDS, which is a separate product from HCP. The HCP search facility This facility is integrated with HCP and works internally to perform searches and return results to the Search Console. The HCP search facility is available only in HCP systems that included it before being upgraded to release 4.1.
Only one of the search facilities can be enabled at any given time. If neither one is enabled, the HCP system does not support searching namespaces. The system associated with the enabled search facility is called the active search system. The active search system (that is, HDDS or HCP) maintains an index of data objects in each search-enabled namespace. This index is based on object content and metadata. The active search system uses the index for fast retrieval of search results. When data objects are added to or removed from the namespace or when object metadata changes, the active search system automatically updates the index to keep it current.
13
HCP hardware
For more information on the HDDS and HCP search facilities, see Administering HCP. For information on using the HCP Search Console, see Searching Namespaces. HCP services HCP services are background processes that each perform a specific function that contributes to the continuous tuning of the HCP system. In particular, services are responsible for optimizing the use of system resources and maintaining the integrity and availability of the stored data. Replication An HCP system can be replicated to another HCP system. Replication, an HCP service, is the process of keeping selected tenants and namespaces in two HCP systems in sync with each other. This entails copying object creations, deletions, and metadata changes, as well as tenant and namespace configurations, from one system to the other. For more information on replication, see Replicating Tenants and Namespaces. Note: Replication is an add-on feature to HCP, with a separate purchase price.
HCP hardware
HCP runs on a networked redundant array of independent nodes (RAIN) or a SAN-attached array of independent nodes (SAIN). SAN stands for storage area network. HCP 300 systems use a RAIN configuration. HCP 500 systems use a SAIN configuration. HCP hardware consists of:
Servers Internal, direct-attached, or SAN-attached storage Networking components such as cables and switches Additional infrastructure items such as racks and power distribution
units (PDUs)
14
HCP hardware
Storage nodes
Storage nodes are the essential part of HCP. They manage the objects that reside in the internal or external storage. To ensure data integrity and continuous availability in case of a hardware or software failure, HCP uses RAID technology and can also store the data and metadata for each object in multiple locations. Each storage node runs the complete HCP software, with the exception of the HCP search facility software, if present. These nodes, therefore, act both as a repository manager and a gateway that enables access to the data in the repository. Runtime operations are distributed among the storage nodes, thereby ensuring reliability and performance as capacity grows. If a node fails, the HCP system adapts by redirecting processing to other nodes, so the stored data remains available to users.
Search nodes
To support the HCP search facility, which is available only in systems that already included it before being upgraded to release 4.1, an HCP system must include one or more search nodes. Search nodes run the HCP search facility software and store the HCP search index in a distributed manner. The portion of the index on each node is protected by RAID storage technology. (The HDDS search facility index is stored in HDDS.) All runtime HCP search facility processing is distributed among the search nodes, thereby ensuring reliability and performance as indexing and querying needs grow. The HDDS search facility does not uses search nodes.
15
HCP hardware
Networking
HCP has both back-end and front-end networks. The isolated back-end network connects the HCP nodes to each other through back-end Ethernet switches. Caution: The back-end network must remain separate from the front-end network. Interfering with the operation of the back-end network or backend switches can cause the HCP system to become inoperable and can result in data loss. Each storage node is configured with two pairs of bonded Ethernet ports that allow external applications to access the system. The recommended setup includes two independent Ethernet switches that connect these ports to the front-end network.
System architecture
The following figures show the architecture of HCP RAIN and SAIN systems and the relationship between two systems involved in replication. RAIN system architecture The figure below shows the architecture of an HCP system that uses internal storage. This system has four storage nodes.
16
HCP hardware
Connects from
Back-end network interface cards (NICs) in each node Front-end NICs in each node Back-end switches Each node Each back-end switch
Connects to
Back-end switches Front-end switches Each other Two PDUs One PDU
SAIN system architecture The figure below shows the architecture of an HCP system that uses external fibre channel storage. This system has four storage nodes and two modular storage trays.
17
HCP hardware
Connects from
Back-end NICs in each node Front-end NICs in each node Back-end switches Each node Each node Each back-end switch Each storage tray
Connects to
Back-end switches Front-end switches Each other External storage Two PDUs One PDU Two PDUs
Notes:
Some HCP SAIN systems include fibre channel switches between the
nodes and the external storage. Network connections for HCP with replication Replication requires a primary HCP system and a secondary HCP system called a replica. The two systems are connected through the front-end network infrastructure.
18
HCP hardware
Logical volume
Node
Multipathing is configured at the storage tier. Zero-copy failover Zero-copy failover is the process of one node in an HCP SAIN system automatically taking over management of storage previously managed by another node that has failed. To support zero-copy failover, each logical volume that stores objects must map to two different storage nodes. The pair of nodes forms a set such that the volumes that map to one of the nodes also map to the other. This is called cross-mapping.
19
HCP software
In the picture below, the red lines represent the physical paths between the two nodes (bottom) and the storage array (top). The broken green lines represent the mappings between each of two sets of logical volumes and each of the two nodes.
Logical volume
Logical volume
Node
Node
Cross-mapping is defined at the storage tier. Zero-copy failover is automatically enabled for an HCP SAIN system during installation.
HCP software
HCP system software consists of an operating system (the Appliance Operating System) and HCP software. For maximum reliability, each storage node in an HCP system runs all the HCP software (except the HCP search facility software, which runs on search nodes). This software includes components that:
Handle object data and metadata Ensure that services run as expected Enable you to configure, monitor, and manage the HCP system and the
tenants and namespaces defined in the system
Support data transfer to and from the HCP system through the HCP
Data Migrator utility
110
111
Software installation activities To install a new HCP system, you need to: 1. Gather the information you need to perform the installation 2. Install the Appliance Operating System (OS) 3. Learn about the HCP software installation program 4. Install the HCP software 5. Verify the installation 6. Perform additional system configuration activities, as needed For more information on these activities, see Chapter 2, Gathering installation information, and Chapter 3, Installing an HCP system. Software upgrade activities To upgrade an existing HCP system from one major or minor release to another, you need to: 1. Gather the information you need to perform the upgrade 2. Upgrade the HCP software (depending on release requirements, this may include the OS) 3. Verify the upgrade If you run into trouble with an upgrade, you may need to restart it. Or, for an offline upgrade, you may need to roll it back. Additionally, you may occasionally need to apply hotfixes to resolve particular customer needs. For more information on these activities, see Chapter 4, Upgrading an HCP system. Node maintenance activities If a node fails, you may need to recover it. The procedure for this differs for storage and search nodes. For information on these activities, see Chapter 6, Maintaining HCP nodes.
112
System hardware reconfiguration activities To accommodate growing and changing HCP systems, you may need to:
Add nodes to the system to support a greater processing load and, for
RAIN systems, increase storage capacity. For information on this activity, see Chapter 5, Adding nodes to an HCP system.
Remove nodes and, for SAIN systems, storage arrays. For information
on this activity, see Chapter 8, Retiring devices. System software reconfiguration activities Occasionally, changes to the HCP configuration become necessary. You may need to:
Change domain name system (DNS) information Change the network configuration Change time settings Change the HCP system serial number Change the HTTP persistent connection timeout interval Enable or disable zero-copy failover Enable or disable DPL one in RAIN systems Change the region count Enable or disable support for the replication feature
For more information on these activities, see Chapter 9, Reconfiguring HCP system software.
113
Service-level system management activities Some HCP management activities are performed by the customer. Others are reserved for service providers. As a service provider, you may need to:
Start, stop, enable, or disable services Reset the checkpoint for a replication link Disable chained writes Access the HCP search facility administrative interface
For more information on these activities, see Chapter 10, Managing an HCP system. Troubleshooting activities Customers can perform some troubleshooting and disaster recovery activities on their own. The ones listed here are reserved for service providers:
Understanding advanced alerts Retrieving the HCP Setup logs Resetting the install user password Resetting a security user password Setting the logging level for the HCP internal logs Downloading the HCP internal logs for a single node Managing encryption keys for log downloads Creating a snapshot of the JVM Using the system service tools Viewing the system status pages Ejecting the CD tray from a node
114
For more information on these activities, see Chapter 11, Troubleshooting an HCP system. Note: As a service provider, you are also responsible for carrying out instructions in service alerts released by the HCP support center.
115
116
2
Gathering installation information
For a new HCP system, the software to be installed consists of an operating system (the Appliance Operating System) and the HCP core software. For a reinstallation of a system that includes search nodes, the software also includes the HCP search facility software. Before you can install the OS and HCP software, you need to know certain information about the system youre creating. For example, you need to know the IP addresses of the HCP nodes, the serial number for the system, and which features the customer wants enabled. The installation information you need comes from both the customer and the HCP distributor. Some of it, such as node IP addresses and serial number, is straightforward. Other information, such as whether to use DNS, requires a customer decision. Most customer decisions are made before the hardware is set up for the HCP system. Once the HCP system is up and running, you may need to do some postinstallation setup. For example, you may need to enable syslog logging or disable the ability to ping the HCP nodes. To do this setup, you need to know what the customer wants. This chapter describes the information you need for installing both the OS and HCP software. It also outlines the information required for any postinstallation setup you may need to do. The descriptions in this chapter are intended to help you with the specifications required by the HCP configuration tool supplied by the HCP distributor. If you dont have such a tool, you can use the forms in Appendix A, Forms, to record the information you collect. You can then keep the completed forms available to facilitate the installation procedure.
21
The physical location. The front-end IP address. The subnet mask for the front-end network. The IP address of the default gateway for the front-end network. This
is the same for all nodes.
22
After the HCP core software is installed, the fourth octet of the back-end IP address cannot change, even if the fourth octet of the front-end IP address changes. Note: Several procedures in this book refer to the highest-numbered storage node. The number assigned to a node is the fourth octet of its back-end IP address. Caution: Once HCP is installed, interfering with the operation of the backend network or back-end switches can cause the system to become inoperable and/or can result in data loss. Front-end network gateway For an OS installation, you need to know the IP address of the default gateway for the front-end network. Ask the customer what IP address to use for this purpose. Once the HCP system is running, you can use the System Management Console to change the gateway IP address you specify during the OS installation. For more information on this, see Changing network settings on page 9-4.
Distributor key
For an HCP core software installation, you need to know the name of the distributor key. This key enables access to the HCP nodes for troubleshooting purposes. For the name of the distributor key, contact the HCP distributor.
23
Networking
When installing the HCP core software, you need this networking information:
The IP address of the gateway for routing requests from the HCP
system to the front-end network. This is the same as the front-end gateway IP address you specify during the OS installation.
24
When installing the core software for an HCP system set up for DNS, you need to know the DNS name of the system and the IP addresses of all the corporate DNS servers. Notes:
You cannot use the DNS name to access an HCP system until the HCP
subdomain is included in the corporate DNS setup.
If the customer doesnt use DNS, you still need provide a DNS name
for the HCP system during the core software installation. This dummy DNS name is used for access to both the Tenant Management Console and the stored data. Once the HCP system is running, you can use the System Management Console to change the DNS name of the system and specify different DNS servers. For more information on this, see Changing DNS settings on page 9-2. For instructions on configuring the HCP subdomain for DNS, see Administering HCP.
Time
The customer needs to decide whether the HCP system should use itself as a time source or should use one or more external time servers. When the time source is internal, the nodes synchronize time among themselves. Because HCP is a closed system, making the time source internal ensures compliance with the applicable regulatory requirements. On the other hand, using internal time can result in clock drift, thereby causing HCP system time to differ from the time settings of other applications in the corporate environment. Resetting the HCP system time to compensate for this drift can affect object retention, thereby violating compliance. External time servers can be corporate or Internet time servers. Using a corporate time server keeps the HCP system time synchronized with other applications in the corporate environment. However, if the corporate time server is not completely secure, compliance is not guaranteed. The same concern about compliance also applies to Internet time servers. Using multiple external time servers helps ensure the integrity of the time source.
25
Tip: Several companies offer time server appliances that enable a system to manage time independently of the concerns mentioned above. Such appliances plug into the HCP network and use a GPS device to keep time. Notes:
The time on the time server is changed by more than 1,000 seconds. The HCP system switches to a different time server whose time differs from the original time server by more than 1,000 seconds.
Internals
For a core software installation, you need to know what the customer wants for several internal settings.
26
Storage configuration During the installation of the HCP core software, you need to specify the type of storage the system uses. The options are internal and external. For SAN-attached storage, you need to specify that the storage is external. For all other types of storage, you need to specify that its internal. Additionally, for SAN-attached storage, you need to know:
Whether any of the storage arrays used by the HCP system are shared
with other applications or systems
Whether any of the storage arrays used by the HCP system are model
WMS-100 or AMS-200
Whether the HCP nodes are blades in one or more Hitachi CB 320
servers and, if so, the IP address of the management module (also called the chassis IP address) for each server Note: If the HCP system is using CB 320 servers, for each server, you need to add the lowest-numbered storage node in the HCP system to the list of SNMP managers in the server configuration. To identify the node, use its front-end IP address. For the community name, use public. If the lowest-numbered node changes at any time (for example, due to prolonged node unavailability or the addition of a node with a lower number), you need to update the configuration of each CB 320 server accordingly. For information on configuring CB 320 servers, see the applicable Hitachi documentation. Serial number When you install HCP, you need to specify the unique five-digit serial number for the system. This number is on a label on the rack that houses the HCP nodes. This label is attached to the side of the rack at the bottom, just inside the left rear door. Once the HCP system is running, you can use the System Management Console to change the serial number (for example, if you mistyped it during installation). For more information on this, see Changing the HCP serial number on page 9-9.
27
Replication If the customer purchased the HCP replication feature, you need to enable the feature when you install the core software. Enabling replication allows the feature to be implemented; it does not create the replication link or start replication activity. Important: Do not enable replication if the customer has not purchased this feature. Doing so makes the system violate the customers license agreement. For information on replication, see Replicating Tenants and Namespaces. Customer support contact information When installing the HCP core software, you have the option of specifying HCP customer support contact information. You specify this information as a text string. After installation, the customer can use the syscontact field in the HCP MIB file (HCP-MIB.txt) to retrieve the contact information through SNMP.
Security
During a core software installation, you can specify that HCP should encrypt all data and metadata stored in the repository. This option, which can be enabled only when the system is installed, ensures the privacy of data and may be required for compliance with certain government regulations. With the repository encrypted, lost or stolen storage devices are useless to parties without the correct encryption key. HCP handles data encryption and decryption automatically, so no human interaction is required. However, once the repository is encrypted, it remains encrypted throughout its lifetime. Encryption cannot be disabled. If you enable encryption for an HCP system, the installation program displays the encryption key while its installing the software and then prompts you to enter the key for confirmation. Before entering the key, write it down on paper. Then give the paper to the customers security administrator for storage in a secure location. Alternatively, if the security administrator is present for the installation, have that person write down and enter the key. This provides the customer with a high level of confidence in the security of their data. Tip: You can use the Encryption Key form on page A-2 to record the encryption key.
28
Important: The encryption key is displayed only once during installation and is not accessible through the HCP System Management Console. Caution: Loss of the encryption key will most likely result in unrecoverable data in the case of catastrophic system failure.
HCP nodes
For an HCP software installation, you need to know the back-end IP address of each HCP node. If youre reinstalling a system that already included search nodes, you also need to know which nodes are storage nodes and which are search nodes. During the installation, you can specify node IP addresses individually or as a single range for each type of node. You can also use a file to supply a list of IP addresses. For information on this last option, see Using a node-list file on page 3-27.
29
Whether to disable the ability for HCP service providers and support
personnel to use SSH to log into the HCP nodes. Important: While disabling SSH provides extra security for the HCP system, it also severely limits the ability of service providers and support personnel to provide support for the HCP system.
Whether to install a new SSL server certificate. In this case, you need
to either have an SSL server certificate to install or have the information needed to generate a certificate signing request.
Whether to enable the HDDS or HCP search facility. Whether to create a replication link (only if the customer has purchased
the replication feature) and, if so, the link properties. Configuration that requires the security role For post-installation configuration activities that require the security role, you need to know:
Whether to enable the HCP Management API Whether to create additional user accounts and, if so, the properties for
each account Configuration that requires either the administrator or security role For post-installation configuration activities that require either the administrator or security role, you need to know:
210
Configuration that requires the service role When HCP is installed, the disposition service is disabled by default. You need to know whether to enable it. Enabling a service requires the service role. Getting more information For more information on:
All of the above options except creating a replication link and enabling
the disposition service, see Administering HCP
Creating a replication link, see Replicating Tenants and Namespaces Enabling the disposition service, see Starting, stopping, enabling, and
disabling services on page 10-3
211
212
3
Installing an HCP system
After the HCP hardware is installed and configured, you need to install the Appliance OS and HCP software on each node. This chapter contains instructions for doing this. Important: This chapter assumes that the HCP system hardware is already installed, configured, and working properly.
31
Check the HCP release notes for any last minute installation
information.
Verify that:
For RAIN systems, the BIOS and RAID group are configured on each node. For instructions on configuring these items, see Maintaining HCP 300 System Hardware. For SAIN systems, the HBA BIOS is enabled and configured on each node. For instructions on loading and configuring the HBA BIOS, see Assembling and Configuring HCP System Hardware.
Note: This configuration is part of the HCP hardware setup, which must be completed before you install the OS.
32
Ensure that the BIOS date in the node is correct. The OS installation
fails if the BIOS date is one or more years in the past.
The formatting of all LUNs used by HCP is 100 percent complete. (When a LUN is completely formatted, its state is Normal.) Each node has full connectivity to the SAN-attached storage. If it doesnt, please contact your HCP support center for help. LUN 0 (zero) is configured for the OS, and sufficient storage is allocated for the data volumes. If these conditions arent true, have the SAN administrator for the customer site reconfigure the storage.
Optionally, for SAIN systems, have the SAN administrator for the
customer site review the storage configuration (see Optionally reviewing the storage configuration below).
33
4. Either press Enter, or let the program default to the install option after 75 seconds. The installation program prompts for the procedure you want to perform.
The console displays the first screenful of storage configuration information for the node:
To page through the rest of the information, use the space bar. To stop the display before youve viewed all the configuration information, press the Q key.
34
3. Power on or reboot the node. The installation program prompts for the installation mode.
4. Either press Enter, or let the program default to the installation option after 75 seconds. The installation program prompts for the procedure you want to perform.
8. When prompted, enter the subnet mask for the front-end IP address or press Enter to accept the default.
9. When prompted, enter the IP address for the front-end network gateway or press Enter to accept the default.
35
The installation program displays the IP addresses and subnet mask you entered and prompts you to confirm them. 11.In response to the confirming prompt:
If the IP addresses and subnet mask are correct, enter y. If the IP addresses or subnet mask are not correct, enter n. In this case, the installation program repeats the prompts, starting again with the front-end IP address.
Tip: After youve successfully entered the IP addresses and subnet mask, you can disconnect the console from the node. You can then use them for the OS installation on another node. The installation program reformats the system volume and installs the OS. This process takes several minutes. While installing the OS, the program reports on its progress. When the installation is complete, the node ejects the CD tray. You can remove the CD from the tray any time after this. When you finish this procedure, the installation program reboots the node. If the node does not reboot automatically, the OS installation failed. In this case, please contact your HCP support center for help.
36
To accept the default, press Enter. To select a different option, type the number or letter you want. Then
press Enter. Menu navigation works like this:
From any menu two or more levels below the main menu, the b option
takes you back up a level.
37
When you return to a menu, the default option is the one following the
one you just executed.
From any menu, the q option exits the wizard without saving your
work.
If the option you select cannot be undone (like the q option), the wizard
prompts twice for confirmation before executing the option. For pictures of all the HCP Setup wizard menus, see Appendix B, HCP Setup wizard menus.
Most of the options in these menus display the currently configured value as the default. When you select a configuration option, the wizard prompts for a value, as in this example:
38
For configuration options that prompt for yes or no, y, yes, n, and no are
all acceptable responses.
The wizard prompts you to confirm each value you specify for a
configuration option.
For certain options, changing the value may change the defaults for
other options. For example, changing the time server configuration from internal to a specific time server changes the Current Date and Time option to not available with time server.
At any time, you can use the v option on the New Install menu to
review the current configuration settings.
39
If the laptop computer is connected to the front-end network, you access the node through its front-end IP address. You can also connect the computer to a back-end switch in the HCP system and access the node through its back-end IP address. For information on this option, see Appendix E, Accessing HCP interfaces through the back-end network. When you use SSH to log into the node, to ensure the continuity of the HCP Setup run, you should open the install shell in a virtual terminal session. The virtual terminal session enables HCP Setup to run even if your SSH session is disconnected. To open the install shell in a virtual terminal session: 1. Using the installation or service SSH key, use SSH to log into the applicable node in the HCP system as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. Note: You need the installation SSH key if the HCP core software is not yet installed. Otherwise, you need the service SSH key. 2. Enter this command to start the virtual terminal session:
screen
Once a virtual terminal session is running, you can detach it from your SSH session. This enables you to shut down your computer while HCP Setup is running. To detach the session: 1. Press Ctrl+A. 2. Press the D key. If your computer shuts down unexpectedly, the virtual terminal session automatically detaches from the SSH session. To reattach a virtual terminal session to an SSH session, enter:
screen -rd
310
Notes:
You can open the install shell directly from the service shell (that is,
without using a virtual terminal session). After starting the HCP installation, you can exit the install shell without affecting the installation process. If you open the install shell again while the installation is still in progress, you return to the same session. However, after the installation process is complete, you cannot return to the session and, therefore, will not be able to see any messages HCP Setup may have issued during the installation.
A keyboard and monitor (or a laptop computer connected to the frontor back-end network)
Check the HCP release notes for any last minute installation
information.
311
For HCP SAIN systems, if any of the nodes in the system do not have
the required version of the HBA firmware, the installation fails.
312
Tip: If you dont see the login prompt, press Enter a few times.
Note: If youre using a laptop computer connected to the front- or back-end network, see Running HCP Setup from a service shell on page 3-9 for login instructions. 2. Insert the HCP core software installation CD into the CD drive in that node. 3. Enter: install A password prompt appears.
4. Enter: Chang3Me! As you type, the characters do not show. Note: If youve previously changed the password for the install user, enter that password instead. Then continue the installation procedure from where you ended the last session. If you havent previously changed the install user password, HCP Setup prompts for the current password again.
313
5. Enter: Chang3Me! The system prompts for a new password for the install user.
6. Enter a new password for the install user. Passwords must follow the standard rules for Unix passwords. In particular, to be valid, a password must include at least one character from two of these three groups: alphabetic, numeric, and special characters. Also, a password cannot be a word found in the dictionary. Changing the install user password in this way is a one-time event. If you need to change it again, use the procedure described in Resetting the install user password on page 11-3. Tip: For the new password, use hcpinsta11, where the last two characters are the number one. 7. When prompted, enter the new password again. The system changes the password and prompts you to continue.
314
3. Do either of these:
CD drive.
2. Enter y or yes. 3. In response to the confirming prompt, enter y or yes to confirm
your entry or n or no to try again. When you enter y or yes, HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
315
HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
5. Press Enter.
To continue the installation without retrieving the HCP search facility installation files:
1. Enter n or no. 2. In response to the confirming prompt, enter y or yes or n or no to
try again. When you enter y or yes, the HCP 4.1 Configuration menu reappears. The HCP 4.1 Configuration menu now includes the option to install HCP.
Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically.
316
When you enter y or yes, the HCP Setup wizard New Install menu appears.
3. From the New Install menu, execute the options for configuring the HCP system. Each option in this menu either opens a lower-level menu with configuration options or leads directly to a configuration option. For pictures of all the HCP Setup wizard menus, see Appendix B, HCP Setup wizard menus. To configure the HCP system:
a. Enter 1 to change the distributor key. b. Enter 2 to change these networking settings:
Whether HCP is (or will be) in the corporate DNS setup DNS name for the HCP system (this is required regardless of whether the customer uses DNS) IP addresses of all the corporate DNS servers
317
Whether the time source is internal or external Current date and time (internal time source only) Time zone Tip: To see a list of valid time zones, enter view for this option. To return to the option from the list, press the Q key.
Whether the HCP storage is internal or external Note: For external storage, you are also prompted to specify:
HCP system serial number Whether to enable replication Important: Do not enable replication if the customer has not purchased this feature. Doing so makes the system violate the customers license agreement.
Contact information for HCP customer support Tip: To specify no contact information, enter a space.
318
g. Enter 7 to identify the nodes in the HCP system. Use the back-end
IP address to identify each node. Tip: If you chose to enter the node IP addresses as literal values, enter the IP address of the lowest-numbered node first. For subsequent IP addresses, HCP Setup presents a default value thats one greater than the previous one you entered.
319
When you enter y or yes, HCP Setup performs a set of installation prechecks and, if they are successful, installs the HCP software on all nodes in the system. This can take from several minutes to several hours, depending on the size of the logical volumes. Important: If you enabled encryption in the system configuration, HCP Setup displays the encryption key after doing some initial setup. It then prompts you to enter the key as you see it. Before entering the encryption key, write it down on paper. You can use the Encryption Key form on page A-5 for this purpose. After you enter the key, HCP Setup proceeds with the installation. You do not get a second chance to see the encryption key, and the key itself is not stored for later retrieval. When the installation is complete, HCP Setup logs you out and reboots the nodes. The console then displays the login prompt. If HCP Setup exits at any time before the installation processing is complete, make a note of all error messages and then contact your HCP support center for help.
your entry or n or no to try again. When you enter y or yes, the wizard returns to the New Install menu, from which you can correct the configuration.
Use the HCP System Management Console. When you use this method,
you can also verify that you correctly entered the HCP system serial number.
Use the HCP Service menu. To use the HCP Service menu, you need a
laptop computer connected to the HCP system front- or back-end network. For more information on this, see Running HCP Setup from a service shell on page 3-9.)
320
If the HCP system is not configured for DNS, enter this URL:
https://node-ip-address:8000 node-ip-address is the front-end IP address of any storage node in the HCP system.
Note: If you inadvertently use http instead of https in the URL, the browser prompts you to open or save a file. Cancel out of the prompt and try again, this time using https. 2. When prompted, accept the self-signed HCP SSL server certificate either permanently or temporarily for this Console session only. You would do the latter, for example, if the customer plans to install a trusted certificate later on. The System Management Console login page appears. If the browser cannot find the System Management Console login page, wait a few minutes; then try again. If the login page still doesnt open after a few minutes, please contact your HCP support center. 3. Check the serial number on the login page. If the serial number is incorrect, you can correct it in the System Management Console. For instructions on doing this, see Changing the HCP serial number on page 9-9. 4. Log into the System Management Console with this username and password: Username: security Password: Chang3Me!
321
The Console displays either the Change Password page or the Nodes page. If the Console displays the Nodes page, it means the nodes are still in the process of starting HCP. This process can take several minutes. When more than half the nodes have completed their startup processing, the Console automatically displays the Change Password page. If the Nodes page remains displayed after several minutes, please contact your HCP support center. 5. On the Change Password page:
In the Existing Password field, type: Chang3Me! In the New Password field, type a new password. Passwords can be up to 64 characters long and can contain any UTF-8 characters, including white space. The minimum length is six characters. To be valid, a password must include at least one character from two of these three groups: alphabetic, numeric, and other. For example, P@sswOrd is a valid password; password is not. Tip: For the new password, use hcpinsta11, where the last two characters are the number one.
In the Confirm New Password field, type your new password again.
Then click on the Update Password button. Important: After successfully verifying an installation, you need to change the password for the security account back to Chang3Me!. 6. In the top-level menu, click on Nodes. 7. On the Nodes page, for each node, check that:
The node status is Available Each logical volume is available and functioning properly (that is, the volume status indicator is )
322
If all the nodes and logical volumes are available, the installation was successful, and the customer can begin creating tenants and namespaces. However, the customer may not want to do this until additional system setup occurs. If any nodes are unavailable or if any logical volumes for available nodes are unavailable (that is, the volume status indicator is either missing or not ), please contact your HCP support center. 8. Do either of these:
Perform additional system configuration, as described in Setting additional configuration options on page 3-26. Do this only if the installation was successful. Change the security user password back to Chang3Me!. Then log out of the System Management Console and close the browser window to ensure that no one can return to the Console without a fresh login.
323
The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
4. From the HCP 4.1 Configuration menu, enter s to display the HCP Service menu. 5. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the HCP Service menu appears.
6. From the HCP Service menu, enter 2 to display the current system status.
324
For each node in the HCP system, the display shows the front-end IP address and the status of the node:
HEALTHY All internal HCP components on the node are running and healthy. UNHEALTHY One or more internal HCP components on the node are not running properly. When the status of a node is UNHEALTHY, the display also shows the status of each internal component. The status of a component can be either AVAILABLE or UNAVAILABLE.
Note: While the HCP system is starting, nodes may show as UNHEALTHY. In this case, wait for the system to finish its startup processing. Then repeat this step to display the current system status again. 7. Review the HCP System Health display. Then press Enter to return to the HCP Service menu. 8. From the HCP Service menu, enter q to return to the HCP 4.1 Configuration menu. 9. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 10.From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 11.In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 12.If any nodes showed as UNHEALTHY after the system completed its startup processing, contact your HCP support center for help.
325
Identifying the nodes in the HCP system by listing them in a file Modifying the HCP configuration file manually for special installation
requirements
326
For these options, you need to perform at least part of the installation as the service user rather than the install user. To log in as the service user, you need to use a laptop computer connected to the HCP system front- or back-end network. For more information on this, see Running HCP Setup from a service shell on page 3-9.)
327
3. Use the chmod command to change the permissions for the node-list file to 644. For example:
chmod 644 /home/service/example/storage_node_list
For more information on virtual terminal sessions, see Running HCP Setup from a service shell on page 3-9. 3. In the virtual terminal session, enter this command to open the HCP install shell:
/home/service/bin/install
4. In response to the prompt to continue, press Enter. The service version of the HCP 4.1 Configuration menu appears.
328
5. If you havent already retrieved the HCP software installation files, do so now. For instructions on doing this, see Step 2: Retrieve the HCP software installation files on page 3-15. If you retrieved the HCP software installation files in a previous installation session, you dont need to retrieve them again. 6. Enter 2 to run the HCP Setup wizard. 7. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the HCP Setup wizard New Install menu appears.
8. From the New Install menu, configure the HCP system, but wait for Step 3: Identify the node-list file below to specify the nodes for which you created a node-list file. For information on using the HCP Setup wizard to configure the system, see HCP Setup program on page 3-7. For configuration instructions, see Step 3: Configure the HCP system on page 3-16.
329
2. Do either of these:
To provide a node-list file for the HCP storage nodes, enter 1. To provide a node-list file for the HCP search nodes, enter 2.
The Storage Node Back-end IP Addresses or Search Node Back-end IP Addresses configuration option appears, as applicable.
3. Enter: file 4. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
330
When you enter y or yes, the wizard prompts for the node-list file.
5. Enter the full path and file name of the node-list file. 6. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the wizard lists the node IP addresses that are in the file and prompts you to confirm that theyre correct.
7. Do either of these:
If the list of IP addresses is correct, enter y or yes. The wizard redisplays the HCP Nodes menu. From here, optionally specify a node-list file for the other type of node. Then continue the installation procedure, as described in Step 4: Execute the installation on page 3-19.
exit
5. Correct the node-list file and repeat this procedure, starting with
331
332
3. Use a text editor such as vi to edit the copy of the cluster.cfg file as instructed by HCP support personnel. 4. Use the chmod command to change the permissions for the copy of the cluster.cfg file to 644. For example:
chmod 644 /home/service/example/cluster.cfg
For more information on virtual terminal sessions, see Running HCP Setup from a service shell on page 3-9. 2. In the virtual terminal session, enter this command to open the HCP install shell:
/home/service/bin/install
333
4. Enter 2 to run the HCP Setup wizard. 5. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the HCP Setup wizard New Install menu appears.
334
2. Enter the full path and file name of the modified configuration file. 3. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the wizard loads the modified configuration file and redisplays the New Install menu. 4. Continue the installation procedure, as described in Step 4: Execute the installation on page 3-19.
335
336
4
Upgrading an HCP system
You can upgrade an HCP system to release 4.1 from a 3.0.1 or later release. You cannot upgrade to release 4.1 directly from a release earlier than 3.0.1. An upgrade can occur with the HCP system online or offline. If an online or offline upgrade does not complete successfully, you can try restarting it. If an offline upgrade still does not complete successfully, you can return the system to its original state by rolling back the upgrade. Occasionally, you may need to apply hotfixes to particular customer installations. Hotfixes are designed to resolve customer problems on a one-off basis. This chapter contains complete instructions for:
Upgrading an HCP system to release 4.1 Restarting an upgrade Rolling back an offline upgrade Installing a hotfix
41
About upgrading
About upgrading
Depending on the releases youre upgrading from and to, an upgrade of an HCP system entails upgrading either both the Appliance OS and the HCP software or only the HCP software:
For an upgrade from a 3.x release or later to release 4.1, you need to
upgrade both the OS and the HCP software.
For an upgrade from release 4.1 to a later 4.x release, you may or may
not need to upgrade the OS along with the HCP software. This depends on the versions of the HCP software youre upgrading from and to. All upgrades apply to both storage and search nodes, if the system includes any. During an upgrade, both the data stored on the storage nodes and the search index stored on the search nodes remain intact. To upgrade an HCP system, you use the same HCP Setup program as you do for a fresh installation. If youre not familiar with HCP Setup, read HCP Setup program on page 3-7 before beginning the upgrade procedure. Normally, you perform the upgrade procedure using a keyboard and monitor connected to the highest-numbered storage node in the HCP system. However, you can also perform the procedure from a laptop computer connected to the front- or back-end network. For more information on this, see Running HCP Setup from a service shell on page 3-9. During an upgrade, you cannot change HCP system configuration options the way you can with a fresh installation. After the upgrade completes, however, you can change certain configuration options from the System Management Console. For more information on what you can change, see Chapter 9, Reconfiguring HCP system software. You cannot upgrade an HCP system while a data migration is in progress. The customer can choose to let the migration finish, in which case, you need to finalize it before you can upgrade the system. Alternatively, the customer can cancel the migration and restart it after the upgrade is complete. For information finalizing a migration, see Finalizing a data migration on page 8-2.
42
If possible, you should perform online upgrades while the load on the
HCP system is light.
During an online upgrade to HCP release 4.1, clients cannot store both
data and custom metadata for an object with a single request. This capability is new with release 4.1 and cannot be used until the upgrade is complete.
43
After the upgrade of the node is finished, clients can remount from that node. If automounter is present on the client, the remount happens automatically.
When you start an online upgrade, all services, except shredding, are
disabled automatically. Those that were enabled before the upgrade started are automatically enabled again after the upgrade is complete. These additional considerations apply to services:
If you enable the protection service while the upgrade is in progress, the upgrade may fail. For a system in which indexing was enabled before the upgrade, objects added to search-enabled namespaces while the upgrade is in progress are not indexed or searchable until after the upgrade is complete.
44
Check the HCP release notes for any relevant last-minute information. Ensure that you have the resources you need, as described in
Resources you need for an upgrade on page 4-6.
45
The install user password for logging into the HCP nodes.
You also need to know whether the upgrade should occur with the system online or offline. For information on making this decision, see Online and offline upgrades on page 4-3.
Conditionally, one Appliance OS installation CD for the new release. You need this only if you need to upgrade the OS along with the HCP software. The HCP core software installation CD. If youre upgrading an HCP system that includes search nodes, the two HCP search facility software installation CDs. You also need these CDs if the system does not yet include any search nodes, but you plan to add one or more after the upgrade.
A keyboard and monitor (or a laptop computer connected to the frontor back-end network)
46
47
What you do To run the upgrade prechecks: 1. Follow the steps in Step 1: Log in for the upgrade on page 4-17 to log into the highest-numbered storage node in the HCP system. 2. Follow the steps in Step 2: Retrieve the files for the upgrade on page 4-19 to retrieve the software installation files for the HCP release youre upgrading to. 3. From the HCP 4.1 Configuration menu, do either of these:
Enter 4 to run the upgrade prechecks for an offline upgrade. Enter 5 to run the upgrade prechecks for an online upgrade.
Tip: If you dont see these options, verify that the Configuration menu you see is for the release youre upgrading to. If the wrong menu is displayed, exit out to the login prompt. Then perform step 1 again, being sure to insert the HCP core software installation CD, as instructed. 4. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes:
If an OS upgrade is required, HCP Setup prompts you to insert the Appliance OS installation CD.
Insert the Appliance OS installation CD into the CD drive in the node youre logged into. Then press Enter. HCP Setup performs the upgrade prechecks, which include checking for available space. Note: Do not remove the OS installation CD from the CD drive until the CD tray is ejected automatically.
If an OS upgrade is not required, HCP Setup performs the upgrade prechecks, which include checking for available space.
48
For a list of all the prechecks HCP Setup performs, see Upgrade prechecks below. 5. In response to the prompt to continue, press Enter. The HCP 4.1 Configuration menu reappears. 6. From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 7. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. Upgrade prechecks When you run the upgrade prechecks, HCP Setup performs only those that are needed to ensure a successful upgrade. For example, if the system doesnt include any search nodes, HCP Setup skips the search package precheck. The table below lists each precheck HCP Setup can perform and indicates what you should do if the precheck fails.
Precheck
all network links
Failure Description
The indicated NIC is not correctly connected to an Ethernet switch. possibly because the cable is not plugged in or is defective. The indicated node is unavailable.
Action
Ensure that the cable is plugged in at both ends and that it is not defective. If the connection appears to be correct, contact your HCP support center for help. Try restarting the node from the System Management Console. If that doesnt resolve the problem, contact your HCP support center for help. Log out of the HCP Configuration menu. Ensure that the correct CD for the upgrade is in the CD drive and that the drive tray is closed. Then log in again and retrieve the files for the upgrade. Have your SAN administrator correct the storage configuration.
correct menu
The version of the HCP Configuration menu currently being used is incorrect for the upgrade.
cross mapping
For SAIN systems with crossmapped nodes, the logical volumes are incorrectly mapped to the indicated node.
49
Precheck
database dump space
Failure Description
For offline upgrades only, insufficient space is available to ensure that a rollback can be performed in the event that the upgrade fails. Insufficient space is available for the upgrade. HCP does not appear to be configured correctly in the corporate DNS, or one or more DNS servers known to HCP are not currently available.
Action
Contact your HCP support center for help.
Free space as described in Freeing space on the HCP nodes below. Reconfigure the corporate DNS to correctly reference HCP, or update the HCP system to identify the correct DNS servers. For information on configuring HCP in the corporate DNS, see Administering HCP. For information on updating HCP DNS settings, see Changing DNS settings on page 9-2.
HBA firmware
The HBA card in one or more nodes does not have the required version of the HBA firmware and/or HBA BIOS.
Either: Enter y to continue the HCP upgrade or prechecks without upgrading the HBA firmware and BIOS. In this case, you need to upgrade the HBA firmware and BIOS, as needed, at a later time. Enter n to stop the HCP upgrade. Then upgrade the HBA firmware and BIOS, as needed, and start the upgrade again.
410
Precheck
HCP is not susceptible to bug 24747
Failure Description
The default namespace has retention classes and either or both of these is true: One or more directories in the namespace are included in an outbound or inbound replication link. The HCP system includes the HCP search facility and the default namespace is search enabled.
Action
Do either of these: To continue the upgrade, enter y. After the upgrade is complete, reassign the applicable retention classes to the affected objects. To cancel the upgrade, enter n.
Under the above conditions, if an object that was already replicated was subsequently assigned to a retention class, HCP did not replicate the change. Similarly, if an object that was already indexed was subsequently assigned to a retention class, HCP did not update the index with the change. If this precheck fails, HCP Setup gives you the option of continuing or canceling the upgrade. log rotate is not running multipath devices The HCP internal logs are currently being rotated. HCP could not communicate with one or more SAN storage arrays. Wait 20 minutes. Then run the prechecks again. Check the physical paths and correct any problems. If all path components appear to be correctly connected and functioning properly, contact your HCP support center for help. Check the physical paths and correct any problems. If all path components appear to be correctly connected and functioning properly, contact your HCP support center for help. End the process if possible. If thats not possible, contact your HCP support center for help.
multipathing
For SAIN systems with multipathed storage, one or both of the physical paths between the indicated node and the storage array are broken, or multipathing is not configured correctly on the node. An internal process is currently accessing /fcfs_*.
411
Precheck
node connections
Failure Description
The indicated node is unavailable.
Action
Try restarting the unavailable node from the System Management Console. If that doesnt resolve the problem, contact your HCP support center for help. Have your SAN administrator correct the search node storage configuration. Perform an offline upgrade instead of an online upgrade.
For SAIN systems with search nodes, the search node storage is configured incorrectly. For online upgrades only, either: The HCP system does not include more than two storage nodes. A previous upgrade attempt was rolled back, and the HCP system includes at least one search node.
run location
The node on which youre running HCP Setup is not the highestnumbered storage node or is not properly configured.
Run HCP Setup from the highestnumbered storage node. If the node youre running on is the highest-numbered storage node, contact your HCP support center for help. Restart HCP Setup as the install user. Execute option 1 from the HCP 4.1 Configuration menu and retrieve both the HCP core software installation files and the HCP search facility installation files. If no user accounts with the service role are enabled, ask the HCP security administrator to enable at least one. If the password for the user account you plan to use to complete the upgrade will expire within seven days, log into the HCP System Management Console using that account and change the account password.
You did not start HCP Setup as the install user. The HCP system includes one or more search nodes, but you did not retrieve the HCP search facility files for the upgrade. All user accounts with the service role either are disabled or will become disabled within seven days. You need an enabled user account with the service role to complete the upgrade.
412
Precheck
smtp batches
Failure Description
More than one thousand batches of emails are waiting to be processed
Action
Disable the SMTP protocol in the Tenant Management Console for the default tenant. Then periodically run the upgrade prechecks until fewer than one thousand batches remain. The amount of time this takes depends on the number of outstanding batches. Try restarting the node from the System Management Console. If that doesnt resolve the problem, contact your HCP support center for help. Contact your HCP support center for help. Have your SAN administrator correct the storage configuration. Change the DNS name for the system to a well-formed DNS name other than hcap-name.domain.com. Contact your HCP support center for help. Contact your HCP support center for help. Check the HCP release notes for the supported upgrade paths. Then get the CDs you need to successfully upgrade the system.
SSH
A problem exists with the ssh keys used by the HCP system. For SAIN systems with zero-copy failover enabled, the indicated node has no standby volumes. The DNS name for the HCP system is either hcap-name.domain.com or not a well-formed DNS name. The indicated node cannot communicate with the other nodes in the HCP system. The HCP system is in a state in which it cannot be upgraded. Upgrades from the currently installed version to the new version are not supported.
system name
systemwide SSH
413
From the HCP distributor, the service SSH key, along with the
passphrase for that key What you do To free space on all the HCP nodes: 1. Ensure that all the nodes in the HCP system are running and healthy. You can use the HCP System Management Console to check this. 2. Insert the HCP core software installation CD into the CD drive in the highest-numbered storage node. Note: If the highest-numbered storage node was added after the system was last installed or upgraded, insert the CD into the storage node that had the highest number at the time of the installation or upgrade. 3. Using the service SSH key, use SSH to log into the node into which you inserted the CD as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 4. Enter this command to open the HCP install shell:
/home/service/bin/install
414
The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
6. From the HCP 4.1 Configuration menu, enter s to display the HCP Service menu. 7. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the HCP Service menu appears.
8. From the HCP Service menu, enter 4 to free space on the HCP nodes.
415
9. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the operation to free space runs. When the operation finishes, a continuation prompt appears.
10.Press Enter to return to the HCP Service menu. 11.From the HCP Service menu, enter q to return to the HCP 4.1 Configuration menu. 12.In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 13.From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 14.In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. If, after this procedure, any nodes still have insufficient space for an upgrade, please contact your HCP support center for help.
416
7. Restore any suspended replication links (see Step 7 (conditional): Resume replication on page 4-27)
417
Tip: If you dont see the login prompt, press Enter a few times.
Note: If youre using a laptop computer connected to the front- or back-end network, see Running HCP Setup from a service shell on page 3-9 for login instructions. 3. Enter: install A password prompt appears.
4. Enter the password for the install user. The system prompts you to continue.
Note: If you dont know the install user password, you can reset it. For instructions on doing this, see Resetting the install user password on page 11-3. 5. Press Enter.
418
The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
Note: If the node you logged into was not the highest-numbered storage node when the HCP system was installed or last upgraded, the HCP 4.1 Configuration menu displays only some of the options shown above.
419
When you enter y or yes, HCP Setup copies the HCP core software installation files from the CD to the node and ejects the CD tray. Note: If HCP Setup cannot copy all the files due to insufficient space, it displays this message:
Wrote only x of y bytes
In this case, you need to free space on the node, as described in Freeing space on the HCP nodes on page 4-14, and then try again to retrieve the installation files. What happens next depends on whether the HCP system includes any search nodes. HCP system with search nodes If the HCP system includes one or more search nodes, HCP Setup prompts for the first search facility software installation CD.
To retrieve the HCP search facility installation files: 1. Insert the first search facility software installation CD into the CD drive and press Enter. HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
2. Insert the second HCP search facility software installation CD into the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
3. Press Enter.
420
The HCP 4.1 Configuration menu appears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically. HCP system without search nodes If the HCP system doesnt include any search nodes, HCP Setup asks whether you want to retrieve the search facility installation files.
Do either of these:
drive.
2. Enter y or yes. 3. In response to the confirming prompt, enter y or yes to confirm your
entry or n or no to try again. When you enter y or yes, HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
5. Press Enter.
421
again. When you enter y or yes, the HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically.
If an OS upgrade is required, HCP Setup prompts you to insert the Appliance OS installation CD.
Insert the Appliance OS installation CD into the CD drive in the node youre logged into. Then press Enter. Note: Do not remove the OS installation CD from the CD drive until the CD tray is ejected automatically.
422
If an OS upgrade is not required, the HCP Setup wizard Upgrade menu appears.
Tip: You can enter s from the Upgrade menu to download and view the current system configuration.
423
HCP Setup performs the upgrade prechecks and, if they are successful, upgrades the software on all nodes in the system. If any of the prechecks fail, HCP Setup exits. In this case, fix the problem and then start the upgrade procedure again. For an online upgrade, HCP Setup reboots each node when the upgrade finishes on it. For an offline upgrade, HCP Setup reboots all the nodes at the same time after the installation finishes. Note: For an offline upgrade, after the nodes reboot, the repository is read-only and the HCP system does not accept configuration changes. The repository automatically becomes read-write and configuration changes are again accepted when you commit the upgrade, as described in Step 6: Complete the upgrade on page 4-25. When the upgrade processing is complete, the login prompt reappears. If HCP Setup exits at any time before the upgrade processing is complete, you can try restarting the upgrade, as described in Restarting an upgrade on page 4-27.
Exit the wizard and contact your HCP support center for help.
424
3. Log into the HCP System Management Console using a user account with the monitor, administrator, or service role. Tip: Youll need the service role for Step 6: Complete the upgrade below. 4. In the top-level menu, click on Nodes. 5. On the Nodes page, for each node, check that:
The node status is Available Each logical volume is available and functioning properly (that is, the volume status indicator is )
If all the nodes and logical volumes are available, the upgrade was successful. Note: Search nodes may take significantly longer than storage nodes to become available. If any of the nodes or logical volumes are unavailable, wait a few minutes and then refresh the page. If any of the nodes or volumes continue to be unavailable, please contact your HCP support center for help. In this case, do not perform Step 6: Complete the upgrade or Step 7 (conditional): Resume replication below. 6. Log out of the System Management Console and close the browser window to ensure that the customer cannot return to the Console without a fresh login.
425
To complete the upgrade of an HCP system: 1. On a client computer, open a web browser window. 2. Clear the web browser cache. 3. Log into the HCP System Management Console with the service role. 4. In the top-level menu, mouse over Configuration to display a secondary menu. 5. In the secondary menu, click on Upgrade. Note: The Upgrade page displays the current and capable internal messaging versions for the HCP system. If these versions are not the same, wait a few minutes and then click on Upgrade again to refresh the page. If the versions are still not the same, contact your HCP support center for help. 6. In the Total number of nodes field on the Upgrade page, type the total number of storage and search nodes in the HCP system. 7. Click on the Commit Upgrade button. A confirming message appears. 8. In the window with the confirming message, click in the checkbox to confirm that you understand the consequences of your action. Then click on the Yes button. The System Management Console displays a message indicating whether the upgrade was successfully completed. If the upgrade was not successfully completed, log out of the System Management Console. Then contact your HCP support center for help. Do not try the upgrade again. Tip: While the upgrade needs to be completed, the alerts on the System Management Console Overview page include:
Click here to complete the upgrade
You see this alert only when logged in as a user with the service role.
426
Restarting an upgrade
9. Log out of the System Management Console and close the browser window to ensure that the customer cannot return to the Console without a fresh login.
Restarting an upgrade
If HCP Setup cannot successfully finish upgrading an HCP system, you can try restarting the upgrade. When you do this, HCP Setup starts upgrading the nodes from the point at which it failed. It does not start over from the beginning. You can restart both online and offline upgrades. You can try restarting an upgrade more than once. However, if the upgrade continues to fail, you should contact your HCP support center for help. When an upgrade fails, HCP Setup displays error messages in the system console and prompts you to press Enter. How long this takes depends on the reason for the failure, the point of failure, and whether the upgrade is being performed with the system online or offline.
427
Restarting an upgrade
When you press Enter, HCP Setup ejects the CD tray and returns to the HCP 4.1 Configuration menu. The menu displays only the options that are relevant for the situation.
Important: Do not reinsert the Appliance Operating System installation CD unless HCP Setup prompts for it during the upgrade restart. To restart an upgrade: 1. From the HCP 4.1 Configuration menu, enter 1 to restart the upgrade. 2. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, HCP Setup restarts the upgrade. When the upgrade is complete, the login prompt reappears. If HCP Setup exits at any time before the upgrade processing is complete, you can try restarting the upgrade again, or you can contact your HCP support center for help. 3. Verify the upgrade, as described in Step 5: Verify the upgrade on page 4-24. 4. Complete the upgrade, as described in Step 6: Complete the upgrade on page 4-25. Important: Do not perform this step with out first performing step 3 above. 5. If you suspended replication on any replication links at the beginning of the upgrade procedure, resume replication on those links. For information on doing this, see Replicating Tenants and Namespaces.
428
429
From the HCP distributor, the service SSH key, along with the
passphrase for that key
430
The HCP 4.1 Configuration menu appears. The menu displays only the options that are relevant for the situation.
3. From the HCP Service menu, enter 8 to roll back the upgrade.
431
Installing a hotfix
4. In response to prompt to continue, enter y or yes to confirm your entry or n or no to try again. The wizard prompts again for confirmation. 5. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes in both step 4 and this step, HCP Setup rolls back the upgrade. When the rollback is complete, HCP Setup reboots each node. If any nodes do not restart automatically, reboot them manually. Important: If HCP Setup exits at any time before the rollback processing is complete, contact your HCP support center for help. Do not try the rollback again, and do not try to restart the upgrade.
Installing a hotfix
A hotfix is an update to an HCP system that resolves a particular problem. Typically, hotfixes are installed only on systems that are experiencing that problem. Each hotfix is made available in a single file. The file name has this format:
hcp-release-number-Hhot-fix-number
For example, the file for hotfix number 1004 for HCP release 4.1 would have this name:
4.1.0-H1004
432
Installing a hotfix
For installation, the hotfix file must be located either on a CD or in a directory on the storage node that had the highest number when HCP was installed or last updated. To copy the file onto the node, you can use an SCP tool such as WinSCP. Tip: Create a directory named hotfix in the /home/service directory on the target node and store the hotfix file in there. When a hotfix is installed, its number and installation date are automatically added to the /opt/arc/hotfixes file on each node in the system. You can check this file to see which hotfixes have been installed on a system.
433
Installing a hotfix
3. In response to the prompt to continue, press Enter. The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
434
Installing a hotfix
3. From the HCP Service menu, enter 5 to install the hotfix. The Install Hotfix menu appears.
4. Do either of these:
435
Installing a hotfix
When the installation is complete, HCP Setup ejects the CD tray and returns to the HCP 4.1 Configuration menu. Note: Do not remove the hotfix installation CD from the CD drive until the CD tray is ejected automatically.
2. Enter the full path and file name of the hotfix file.
HCP Setup installs the hotfix. When the installation is complete, HCP Setup returns to the HCP 4.1 Configuration menu. 5. From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 6. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
436
5
Adding nodes to an HCP system
You can add nodes to an HCP system to expand its processing power or to increase its storage capacity. You can add both storage and search nodes. The procedure for adding nodes to an HCP system has both a hardware portion and a software portion. The hardware portion entails racking the new nodes and physically connecting them to the rest of the system. The software portion entails installing both the Appliance OS and the HCP software on the new nodes. This chapter contains instructions for the software portion of adding nodes to an HCP system. For the hardware instructions, see the applicable HCP hardware maintenance documentation.
51
52
If DNS is in use at the customer site, the customer may want to add any new storage nodes as master servers in the HCP stub zone in the corporate DNS setup. For DNS configuration instructions, see Administering HCP. Notes:
For SAIN systems that use cross-mapped storage, you need to add
both nodes in a cross-mapped pair in the same node addition operation.
53
Add the lowest-numbered storage node in the HCP system to the list of
SNMP managers in the configuration of the new CB 320 server. To identify the node, use its front-end IP address. For the community name, use public. If any of the nodes youre adding has a lower number than the current lowest-numbered storage node in the HCP system, you need to update the list of SNMP managers in the configuration of each CB 320 server already included in the system. For information on configuring CB 320 servers, see the applicable Hitachi documentation.
Ensure that the HCP system is running and healthy. You can use the
HCP System Management Console to check this.
Check the HCP release notes for any relevant last-minute information. Complete the HCP Node Addition Information form on page A-6 so the
information you need for the node addition is readily available (see Information you need for a node addition below).
54
For information on the front-end and back-end IP addresses for new nodes, see IP addresses for new nodes on page 5-7.
The install user password for logging into the HCP nodes The version numbers of the currently installed Appliance OS and HCP
software Use the HCP Node Addition Information form on page A-6 to record this information so its readily available when you perform the node addition procedure.
55
One or more of the Appliance OS installation CDs that match the version of the currently installed Appliance OS. Tip: For the quickest installation, have one CD for each new node. That way, you can install the OS on multiple nodes at the same time.
Conditionally, the HCP software installation CDs that match the version of the currently installed HCP software:
If the highest-numbered existing storage node in the HCP system is not the one that had the highest number when the system was installed or last upgraded (which can happen if storage nodes have previously been added to the system), and:
The HCP system doesnt include any search nodes and youre not adding any search nodes, you need only the HCP core software installation CD. The HCP system already has one or more search nodes and/ or youre adding one or more search nodes, you need both the HCP core software and the HCP search facility software installation CDs.
If the highest-numbered existing storage node in the HCP system is the one that had the highest number when the system was installed or last upgraded, you need both the HCP core software and the HCP search facility software installation CDs if all of these conditions apply:
The HCP system doesnt include any search nodes. Youre adding one or more search nodes. You did not retrieve the search facility installation files when the HCP software was installed or last upgraded.
Otherwise, you dont need any of the HCP software installation CDs.
56
A keyboard and monitor (or a laptop computer connected to the frontor back-end network)
The IP address for the front-end network gateway must be the same as
on the existing nodes.
57
2. Press Enter. The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
Note: If the node you logged into was not the highest-numbered storage node when the HCP system was installed or last upgraded, the HCP 4.1 Configuration menu displays only some of the options shown above.
58
To retrieve the HCP search facility installation files: 1. Insert the first search facility software installation CD into the CD drive and press Enter. HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
2. Insert the second HCP search facility software installation CD into the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
59
3. Press Enter. The HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically. HCP system without search nodes If the HCP system doesnt include any search nodes, HCP Setup asks whether you want to retrieve the search facility installation files.
Do either of these:
drive.
2. Enter y or yes. 3. In response to the confirming prompt, enter y or yes to confirm your
entry or n or no to try again. When you enter y or yes, HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
510
again. When you enter y or yes, the HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically.
3. If youre adding one or more storage nodes, from the Membership Update menu, enter 1. Then follow the on-screen instructions to identify each node.
511
4. If youre adding one or more search nodes, from the Membership Update menu, enter 2. Then follow the on-screen instructions to identify each node. Be sure to use the back-end IP address to identify each node. Tip: After you identify the nodes to be added, you can enter v from the Membership Update menu to download and view the current system configuration.
2. In response to the confirming prompt, enter y or yes to confirm that you want to perform the procedure or n or no to back out. The wizard prompts again for confirmation. 3. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes in both step 2 and this step, the wizard downloads and displays the current system configuration. 4. Review the configuration.
512
5. Do either of these:
HCP Setup performs a series of prechecks and, if they are successful, installs the HCP software on all the new nodes. If any of the prechecks fail, HCP Setup exits. In this case, fix the problem and then start the node-addition procedure again. When the node addition is complete, HCP Setup reboots the new nodes, and the HCP 4.1 Configuration menu reappears. If HCP Setup exits at any time before the node addition processing is complete, please contact your HCP support center for help.
6. From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 7. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
513
The node status is Available Each logical volume is available and functioning properly (that is, the volume status indicator is )
If all the new nodes and logical volumes are available (which may take a few minutes), the node addition was successful. Note: Search nodes may take significantly longer than storage nodes to become available. If any of the new nodes or logical volumes are unavailable, wait a few minutes and then refresh the page. If any of the new nodes or volumes continue to be unavailable, please contact your HCP support center for help. 5. Log out of the System Management Console and close the browser window to ensure that the customer cannot return to the Console without a fresh login.
514
6
Maintaining HCP nodes
As a service provider, you are responsible for maintaining the nodes in an HCP system. You may need to:
61
If the OS file system is corrupt, you need to recover the OS. This
leaves the data file systems intact, so no data is lost. For this procedure, see Before you begin a storage node recovery below and Recovering the OS on a storage node on page 6-5. If zero-copy failover is in effect, recovering the OS enables the node to take back management of its failed-over logical volumes. For information on zero-copy failover, see Zero-copy failover on page 1-9. Caution: In HCP RAIN systems, do not use the OS recovery procedure in this chapter to recover a node in which the OS volume is not separate from the data volumes. Doing so can cause permanent data loss in namespaces with a DPL of one.
If the data file systems are corrupt. you need to recover all the logical
volumes on the node. This destroys all stored data in the volumes managed by the node. For this procedure, see Before you begin a storage node recovery below and Recovering all the logical volumes on a storage node on page 6-12. In namespaces with a DPL of two or higher, the protection service ensures that copies of all the objects stored on the node exist on other nodes as well. As a result, no data in those namespaces is lost during this recovery operation. Caution: Recovering all the logical volumes on a node can cause permanent data loss in namespaces with a DPL of one. Important: The node recovery procedures described in this section are for recovering one or more storage nodes in an otherwise healthy HCP system. Do not use either of these procedures to recover from a failed upgrade. In the case of a failed upgrade, restart the upgrade, as described in Restarting an upgrade on page 4-27.
62
Check the HCP release notes for any relevant last-minute information. Ensure that at least one other storage node in the HCP system is
running and healthy.
Ensure that the BIOS time on each node to be recovered is within a few
minutes of the HCP system time.
For the HCP software portion of each procedure, you use SSH to log in
from a laptop computer.
The physical location of each node The front-end IP address of each node The subnet mask for the front-end network The IP address of the front-end network gateway The back-end IP address of each node
63
Internally, HCP uses a database for metadata storage. When recovering the OS, you have the option of recovering this database as well. Before you begin the OS recovery procedure, you need to know whether or not to recover the database. Database recovery is possible only when more than half the nodes in the HCP system are running and healthy.
One Appliance OS installation CD that matches the version of the currently installed Appliance OS. Tip: For the quickest node recovery time, have one CD for each node you want to recover. That way, you can install the OS on multiple nodes at the same time.
Conditionally, the HCP core software installation CD that matches the version of the currently installed HCP software. You need this only if any of these is true:
The highest-numbered storage node in the HCP system is not the one that had the highest number when the system was installed or last upgraded. This can happen if nodes have been added to the system. The highest-numbered storage node is not running and healthy. You have already retrieved the core software installation files for a later version of HCP. Typically, this would be in preparation for an upgrade of the HCP system.
If the HCP system includes any search nodes and you need the HCP core software installation CD, you also need the HCP search facility software installation CDs.
The file containing the service SSH key for the HCP system, along with the passphrase for that key.
64
65
Caution: When reinstalling the OS, be sure to use the same front- and back-end IP addresses as the node had before. Changing the IP addresses can cause the HCP system to become inoperable.
66
The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
Note: If the node you logged into was not the highest-numbered storage node when the HCP system was installed or last upgraded, the HCP 4.1 Configuration menu displays only some of the options shown above.
67
HCP system with search nodes If the HCP system includes one or more search nodes, HCP Setup prompts for the first search facility software installation CD.
To retrieve the HCP search facility installation files: 1. Insert the first search facility software installation CD into the CD drive and press Enter. HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
2. Insert the second HCP search facility software installation CD into the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
3. Press Enter. The HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically. HCP system without search nodes If the HCP system doesnt include any search nodes, HCP Setup asks whether you want to retrieve the search facility installation files.
68
Do either of these:
drive.
2. Enter y or yes. 3. In response to the confirming prompt, enter y or yes to confirm your
entry or n or no to try again. When you enter y or yes, HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
5. Press Enter.
again. When you enter y or yes, the HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically.
69
3. From the HCP Service menu, enter 1 for recovery operations. 4. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, HCP Setup displays the Node Recovery menu.
610
To recover selected storage nodes, enter 1. Then follow the onscreen instructions to identify each node you want to recover. If you identify fewer than half the nodes in the HCP system, HCP Setup asks whether you want to recover the database on those nodes. In response:
1. Enter y or yes to recover the database along with the OS or n or
your entry or n or no to try again. When you enter y or yes, HCP Setup displays a unique key and prompts you to enter it back. If you identify half or more of the nodes in the HCP system, HCP Setup displays a unique key and prompts you to enter it back.
To recover all storage nodes, enter 3. HCP Setup displays a unique key and prompts you to enter it back.
6. Enter the unique key exactly as it is shown. HCP Setup performs a series of prechecks and, if they are successful, recovers the OS on the selected nodes or all nodes, as applicable. If any of the prechecks fail, HCP Setup exits. In this case, fix the problem and then start the OS recovery procedure again. When the node recovery is complete, the HCP Service menu reappears. Important: If HCP Setup exits at any time before the OS recovery processing is complete, please contact your HCP support center for help. Do not try the OS recovery again. 7. From the HCP Service menu, enter q to return to the HCP 4.1 Configuration menu. 8. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 9. From the HCP 4.1 Configuration menu, enter q to log out of the install shell.
611
10.In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
612
3. Using the service SSH key, use SSH to log into the highest-numbered healthy storage node as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 4. Enter this command to set the FORCE_FORMAT environment variable:
export FORCE_FORMAT=1
2. In response to the prompt to continue, press Enter. The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
Note: If the node you logged into was not the highest-numbered storage node when the HCP system was installed or last upgraded, the HCP 4.1 Configuration menu displays only some of the options shown above.
613
Step 3 (conditional): Retrieve the installation files for the logical-volume recovery
If you need to retrieve the HCP software installation files, as described in Resources you need for a storage node recovery on page 6-4: 1. Insert the HCP core software installation CD into the CD drive on the node youre logged into. 2. From the HCP 4.1 Configuration menu, enter 1 (one) to retrieve the installation files from the CD. 3. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, HCP Setup copies the core software installation files from the CD to the node and ejects the CD tray. What happens next depends on whether the HCP system includes any search nodes. HCP system with search nodes If the HCP system includes one or more search nodes, HCP Setup prompts for the first search facility software installation CD.
To retrieve the HCP search facility installation files: 1. Insert the first search facility software installation CD into the CD drive and press Enter. HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
2. Insert the second HCP search facility software installation CD into the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
614
3. Press Enter. The HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically. HCP system without search nodes If the HCP system doesnt include any search nodes, HCP Setup asks whether you want to retrieve the search facility installation files.
Do either of these:
drive.
2. Enter y or yes. 3. In response to the confirming prompt, enter y or yes to confirm your
entry or n or no to try again. When you enter y or yes, HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
5. Press Enter.
615
again. When you enter y or yes, the HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically.
616
4. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, HCP Setup displays the Node Recovery menu.
5. From the Node Recovery menu, enter 1 to recover selected storage nodes. Then follow the on-screen instructions to identify each node you want to recover. HCP Setup displays a unique key and prompts you to enter it back. 6. Enter the unique key exactly as it is shown. HCP Setup performs a series of prechecks and, if they are successful, recovers all the logical volumes on the selected nodes or all nodes, as applicable. If any of the prechecks fail, HCP Setup exits. In this case, fix the problem and then start the logical-volume recovery procedure again. When the node recovery is complete, the HCP Service menu reappears. Important: If HCP Setup exits at any time before the logical-volume recovery processing is complete, please contact your HCP support center for help. Do not try the logical-volume recovery again. 7. From the HCP Service menu, enter q to return to the HCP 4.1 Configuration menu. 8. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 9. From the HCP 4.1 Configuration menu, enter q to log out of the install shell.
617
10.In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 11.Reenable zero-copy failover on the HCP system. For information on doing this, see Enabling or disabling zero-copy failover on page 9-10. Executing the logical-volume recovery on all nodes To execute the logical-volume recovery on all the nodes in the HCP system, you need to perform a fresh installation of the HCP software. For instructions on doing this, see Step 3: Configure the HCP system on page 3-16 and Step 4: Execute the installation on page 3-19.
The node status is Available Each logical volume is available and functioning properly (that is, the volume status indicator is )
If all the recovered nodes and all their logical volumes are available, the node recovery was successful. If any of the recovered nodes or any of their logical volumes are unavailable, wait a few minutes and then refresh the page. If any of the nodes or volumes continue to be unavailable, please contact your HCP support center for help. 4. Log out of the System Management Console and close the browser window to ensure that the customer cannot return to the Console without a fresh login.
618
619
One Appliance OS installation CD that matches the version of the currently installed Appliance OS. Tip: For the quickest node recovery time, have one CD for each node you want to recover. That way, you can install the OS on multiple nodes at the same time.
Conditionally, the set of HCP core software and HCP search facility software installation CDs that matches the version of the currently installed HCP software. You need this only if any of these is true:
The highest-numbered storage node in the HCP system is not the one that had the highest number when the system was installed or last upgraded. This can happen if nodes have been added to the system. You have already retrieved the HCP software installation files for a later version of HCP. Typically, this would be in preparation for an upgrade of the HCP system.
The file containing the service SSH key for the HCP system, along with the passphrase for that key.
620
621
The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
4. Insert the first HCP search facility software installation CD into the CD drive and press Enter.
622
HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
5. Insert the second HCP search facility software installation CD into the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
6. Press Enter. The HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically.
623
3. From the HCP Service menu, enter 1 for recovery operations. 4. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, HCP Setup displays the Node Recovery menu.
5. From the Node Recovery menu, enter 2 to recover all search nodes. HCP Setup displays a unique key and prompts you to enter it back. 6. Enter the unique key exactly as it is shown. HCP Setup performs a series of prechecks and, if they are successful, redeploys the HCP software on all the search nodes in the system. If any of the prechecks fail, HCP Setup exits. In this case, fix the problem and then start the search node recovery procedure again.
624
When this process is complete, the HCP Service menu reappears. 7. From the HCP Service menu, enter q to return to the HCP 4.1 Configuration menu. 8. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 9. From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 10.In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
The node status is Available Each logical volume is available and functioning properly (that is, the volume status indicator is ( )
If all the search nodes and all their logical volumes are available, the node recovery was successful. If any of the search nodes or their logical volumes are unavailable, wait a few minutes and then refresh the page. If any of the nodes or volumes continue to be unavailable, please contact your HCP support center for help. 4. Log out of the System Management Console and close the browser window to ensure that the customer cannot return to the Console without a fresh login.
625
626
7
Managing storage in an HCP SAIN system
To accommodate a growing repository, you can add storage to an HCP system. To add storage to a RAIN system, you need to add storage nodes. To add storage to a SAIN system, you can:
Add storage nodes Add logical volumes to existing storage nodes Replace an existing storage array with a larger array
Customers may also choose to replace an existing array for other reasons (for example, if they are repurposing the array). When replacing an existing storage array, you need to migrate the data from the old array to the new array. Additionally, customers may choose to virtualize an existing storage array behind an enterprise storage array without moving the data. In this case, the customer gains the storage management features associated with the enterprise array while keeping the data in place. This chapter contains instructions for:
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
71
Check the HCP release notes for any relevant last-minute information. Ensure that you have the information you need, as described in
Information you need for a storage addition below.
72
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
A keyboard and monitor (or a laptop computer connected to the frontor back-end network).
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
73
4. Conditionally, retrieve the HCP software installation files (see Step 4 (conditional): Retrieve the HCP software installation files on page 7-5). 5. Make the new logical volumes known to HCP (see Step 5: Make the new logical volumes known to HCP on page 7-8). 6. Complete the storage addition (see Step 6: Complete the storage addition on page 7-10).
74
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
Note: If the node you logged into was not the highest-numbered storage node when the HCP system was installed or last upgraded, the HCP 4.1 Configuration menu displays only some of the options shown above.
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
75
HCP system with search nodes If the HCP system includes one or more search nodes, HCP Setup prompts for the first search facility software installation CD.
To retrieve the HCP search facility installation files: 1. Insert the first search facility software installation CD into the CD drive and press Enter. HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
2. Insert the second HCP search facility software installation CD into the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
3. Press Enter. The HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically. HCP system without search nodes If the HCP system doesnt include any search nodes, HCP Setup asks whether you want to retrieve the search facility installation files.
76
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
Do either of these:
drive.
2. Enter y or yes. 3. In response to the confirming prompt, enter y or yes to confirm your
entry or n or no to try again. When you enter y or yes, HCP Setup copies the first set of search facility installation files from the CD to the node. When this is done, HCP Setup ejects the CD tray and prompts for the second search facility software installation CD.
the CD drive and press Enter. HCP Setup copies the second set of search facility installation files from the CD to the node. When this is done, HCP Setup prompts you to continue.
5. Press Enter.
again. When you enter y or yes, the HCP 4.1 Configuration menu reappears. Note: Do not remove the last installation CD from the CD drive until the CD tray is ejected automatically.
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
77
3. Enter u to perform an online storage addition or o to perform an offline storage addition. 4. In response to the prompt to continue, enter y or yes to continue or n or no to stop the procedure. Note: If you choose to stop the procedure, remember to reenable the protection service if it was enabled before you started. For instructions on doing this, see Managing HCP services on page 10-2. 5. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes in both step 4 and this step, HCP Setup performs a series of prechecks, including a check of the new LUN mappings and of whether the new LUNs are currently reserved by any other systems. If the prechecks are successful, HCP lists the new logical volumes and asks whether the list is correct.
78
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
Do either of these:
Caution: Do not enter y or yes if the list of logical volumes is not correct. Doing so carries the risk of destroying data in the repository and/or leaving the system inoperable.
2. In response to the confirming prompt, enter y or yes to confirm
your entry or n or no to try again. When you enter y or yes, HCP Setup formats the new logical volumes and then prompts you to press Enter to continue. Note: If HCP Setup exits unexpectedly while its formatting the new logical volumes, please contact your HCP support center for help.
3. Press Enter.
install shell.
5. In response to the confirming prompt, enter y or yes to confirm
your entry or n or no to try again. When you enter y or yes, HCP Setup prompts you to press Enter to continue.
3. Press Enter.
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
79
If the prechecks are unsuccessful, HCP reports the errors it found and prompts you to press Enter to continue:
a. Press Enter.
storage configuration, tell the SAN administrator to reconfigure the logical volumes at the storage tier. Then start this procedure again.
The node status is Available Each logical volume (old and new) is available and functioning properly (that is, the volume status indicator is )
If all the nodes and logical volumes are available, the storage addition was successful. If any of the nodes or logical volumes are unavailable, wait a few minutes and then refresh the page. If any of the nodes or volumes continue to be unavailable, please contact your HCP support center for help.
710
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
4. Log out of the System Management Console and close the browser window to ensure that the customer cannot return to the Console without a fresh login.
Upgrade an existing storage array to a larger model. Repurpose an existing storage array, thereby requiring a new or
different array for the HCP system.
Use HDS TrueCopy to migrate the data directly between the existing
array and the new or different array.
Virtualize the existing storage array behind an enterprise array and use
HDS Tiered Storage Manager to migrate the data to the enterprise array.
Virtualize the existing storage array and the new or different storage
array behind an enterprise array and use HDS Tiered Storage Manager to migrate the data to the new storage array.
Have the customer use the data migration service to migrate the data
off the node. Then finalize the migration, as described in Finalizing a data migration on page 8-2. Both TrueCopy and Tiered Storage Manager perform block-level replication of storage from one array to another.
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
711
All four migration methods listed above have brief periods during which the HCP system is unavailable. So does virtualizing storage without migrating the data. Important: The following procedures assume you are familiar with HCP 500 hardware, including the nodes, switches, and storage arrays. If you are not, please contact your HCP support center for help before performing any of these procedures.
From the HCP distributor, one Appliance OS installation CD that matches the version of the currently installed Appliance OS A keyboard and monitor
This step leaves the nodes powered off. 5. Power all the nodes back on (see Step 5: Power on the nodes on page 7-15). 6. Reenable zero-copy failover on the HCP system (see Step 6: Reenable zero-copy failover on page 7-15).
712
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
713
to enter the HBA BIOS configuration. 3. In the HBA BIOS configuration, reconfigure both paths to point to the new boot LUN by WWID. To do this, for each path:
a. Delete the existing boot LUN. b. Scan for the new boot LUN.
If the node cannot find the new boot LUN, verify that the new storage array and the SAN fabric are correctly configured. If the configuration is correct, contact your HCP support center for help. 4. Insert the Appliance OS installation CD into the CD drive in the node. 5. Reboot the node. The installation program prompts for the installation mode.
714
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
You have 75 seconds to do this. If you wait too long, the node continues to boot from the CD. In this case, press Ctrl+Alt+Delete to reboot the node and start over.
san_update is a script that identifies the WWIDs for the new LUNs,
updates configuration files on the node, and builds a new boot image. When the script finishes processing, it automatically ejects the CD tray and shuts down the node. Note: Do not remove the OS CD from the CD drive until the CD tray is ejected automatically.
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
715
3. Update the HCP nodes to recognize the new array (see Step 4: Update the HCP nodes to recognize the new array on page 7-14). For this step, you need:
From the HCP distributor, one Appliance OS installation CD that matches the version of the currently installed Appliance OS A keyboard and monitor
4. Reenable zero-copy failover on the HCP system (see Step 6: Reenable zero-copy failover above).
Fixed speed (not auto-negotiate) Fabric disabled Mode FC-AL Host type Windows Extended (C2)
If LUN security is required, add the WWPNs from the enterprise array to the applicable default host group. If LUN security is not required, disable it on the ports used for the virtualization links.
716
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
Reconfiguring HCP to use a new or different storage array b. Map the LUNs to be virtualized to the default host group on the
applicable ports. Note: If the ports are not dedicated, you need to create a new host group for this purpose instead of using the default host group. 5. On the enterprise array:
a. Create a path group consisting of the ports on the existing array. b. Map the LUNs to the LUNs on the existing array, specifying the
external volume number and control unit number as part of the process. If possible, use the same LUN numbers on the enterprise array as on the existing array, but use a unique control unit number.
c. Create host groups on the ports to which the HCP nodes will connect
through the SAN fabrics. Make these host groups the same as the host groups that use the equivalent LUNs on the existing array.
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
717
718
Managing storage in an HCP SAIN system Installing and Maintaining an HCP System
8
Retiring devices
To properly retire a node in an HCP RAIN or a storage array in a SAIN system, you do not simply unplug the device from the system. For a device that stores objects, you first need to move the data off the device to another location in the system. For both storage and search nodes, you also need to shut down the HCP software on the node in an orderly fashion. The normal way to retire a storage node in an HCP RAIN system is to finalize a data migration that was started by an HCP system administrator. Rarely, however, you may need to retire a node by removing it from the system without first migrating the data off it. For SAIN systems, the normal way to retire a storage array, is to finalize a data migration that was started by an HCP system administrator. You dont normally retire storage nodes in a SAIN system; instead, you replace them. For both RAIN and SAIN systems, you retire a search node by removing it from the system without first migrating the data off it. This chapter contains instructions for finalizing a data migration and for removing a node without first migrating the data off it.
81
82
Because HCP is shut down only on the nodes being retired, the system remains available to users and applications while the data migration is being finalized. After a data migration is finalized, the Nodes page in the HCP System Management Console displays Removed in the controls column for the retired nodes. Finalizing a migration in a SAIN system When you finalize a data migration in a SAIN system, HCP Setup: 1. Shuts down HCP on all nodes in the HCP system. 2. Migrates the system-level data off the storage arrays that are being retired. 3. Removes the data and OS LUNs on the selected arrays from the HCP configuration. 4. Powers off all nodes in the system. If any nodes do not power off automatically, you need to power them off manually. For information on powering off nodes, see Finalizing a migration in a RAIN system above. Because HCP is shut down on all nodes during this procedure, the system is unavailable to users and applications while the data migration is being finalized. After finalizing the migration, you need to update the fibre channel HBA card BIOS in each node from which the OS LUN was retired so that the node boots off the new OS LUN. Additionally, if any search nodes had LUNs on the retired arrays, the SAN storage administrator for the site needs to map new LUNs to those nodes. Then, if the HCP search facility is enabled (or when it is subsequently enabled), you need to rebuild the search index.
Check the HCP release notes for any relevant last-minute information. Ensure that the status of the data migration in the HCP System
Management Console is Migrated.
83
Ensure that you have the information and resources you need, as
described below.
From the HCP distributor, the file containing the service SSH key for the
HCP system, along with the passphrase for that key.
84
3. In response to the prompt to continue, press Enter. The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
Note: If the node you logged into was not the highest-numbered storage node when the HCP system was installed or last upgraded, the HCP 4.1 Configuration menu displays only some of the options shown above.
85
3. From the HCP Service menu, enter 6 for the migration finalization. The wizard prompts you to continue.
4. Enter y or yes to confirm your entry or n or no to try again. 5. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
86
When you enter y or yes in step 4 and this step, HCP Setup performs a series of prechecks, including checking that the data migration is complete and, for SAIN systems, that all the nodes have the required version of the fibre channel HBA firmware. In the latter case, you can continue with the migration finalization. However, after the migration is finalized, the firmware must be upgraded to enable the nodes to operate correctly. If all the other prechecks are successful, HCP Setup finalizes the data migration. If any of the prechecks fail, HCP Setup exits. In this case, fix the problem and then start the upgrade procedure again. If HCP Setup exits at any time before it finishes finalizing the migration, please contact your HCP support center for help. 6. For a RAIN system, in response to the prompt to continue, press Enter. The HCP 4.1 Configuration menu reappears. 7. From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 8. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
87
arrays.
2. For modular arrays housed in HCP system racks: a. Disconnect both ends of the cables that connect the trays of
Step 4 (SAIN systems only; conditional): Reinstall the Appliance OS on the affected search nodes
For a SAIN System, if any search nodes had LUNs from the retired arrays: 1. Have the SAN storage administrator for the site map new LUNs to those nodes. 2. Install the Appliance OS on those nodes. For instructions on this, see Installing the Appliance Operating System on page 3-2.
Step 5 (SAIN systems only; conditional): Update the fibre channel HBA firmware and reboot the nodes
After finalizing a data migration in a SAIN system, for each node:
If the migration finalization prechecks indicated that the node does not
have the required version of the fibre channel HBA firmware, upgrade the HBA firmware in the node.
If the OS LUN for the node was on a retired array, update the fibre
channel HBA BIOS in the node so that the node boots off the new OS LUN. Then reboot the nodes one at a time. For information on configuring the fibre channel HBA firmware, see Assembling and Configuring HCP System Hardware.
88
Removing nodes
To redeploy the HCP software on the search nodes, you use the procedure for adding nodes to the HCP system, specifying the existing search nodes as the nodes to be added. For information on this procedure, see Chapter 5, Adding nodes to an HCP system. Redeploying the software on the search nodes causes whats left of the existing index to be deleted. While the index is being rebuilt, searches may return incomplete results. The amount of time required to rebuild the search index depends on the number and size of the objects to be indexed. With a very large amount of data, this process can take several days.
Removing nodes
Very rarely, you may need to remove a storage or search node from an HCP system. This can happen, for example, if you need the node IP address for some other system. You cannot use the HCP System Management Console to remove a node while a data migration is in progress, even if that node is not a source for the migration. Caution: Do not remove a storage node from an HCP SAIN system unless all data has been migrated off all storage arrays with LUNs mapped to that node. Doing so can result in data loss. When you remove the node after the data migration has been finalized, also remove the node with which that node forms a cross-mapped pair. About removing storage nodes Removing a storage node tells the HCP system that the back-end IP address of the node is no longer part of the system. Node removal is not part of the normal procedure for replacing a storage node that has failed or for upgrading to newer hardware. In these cases, the new node can use the same back-end IP address as the one being replaced. You cannot remove a storage node if doing so would result in fewer than half of the nodes in the system being available. For example, if the HCP system includes four storage nodes but only two of the nodes are available, you cannot remove either of the two available nodes.
89
Removing nodes
After you remove a storage node, the Nodes page in the HCP System Management Console displays Removed in the controls column for the retired nodes. Important:
Once you remove a storage node from the HCP system, the system can
never again include a node with the same back-end IP address as that node. If you remove a search node, the back-end IP address of that node can be reused.
810
Removing nodes
The System Management Console displays the Nodes page, which shows the list of nodes in the HCP system. 3. If the node you want to remove is currently running, click on the shutdown control ( ) for it. A confirming message appears. 4. In the Reason field in the message window, type the reason why youre shutting down the node. This text can be up to 1,024 characters long and can contain any UTF-8 characters, including white space. 5. Click on the Shut Down Node button in the message window. HCP shuts down the node. 6. Click on the remove control ( ) for the node. This control appears only while the node is unavailable. Tip: You can also use buttons at the bottom of the Storage/Search Node page to shut down and remove a node. 7. In response to the confirming message, click on the Remove Node button. HCP removes the node from the system. 8. If the node you removed was a search node and:
The system still includes at least one other search node, add any one such node to the system as if it were a new node. For instructions on doing this, see Chapter 5, Adding nodes to an HCP system. It was the last search node in the system:
1. In the top-level menu in the System Management Console,
811
Removing nodes
812
9
Reconfiguring HCP system software
You can reconfigure HCP system software in several ways:
Adding and removing chassis IP addresses in an HCP SAIN system that uses CB 320 servers Changing the HCP serial number Changing the HTTP persistent connection timeout interval Enabling or disabling zero-copy failover Enabling or disabling DPL one in RAIN systems Changing the region count Enabling the replication feature
91
You perform most system software reconfiguration activities from the HCP System Management Console. Each of those activities requires a user account with the service role. To enable or disable DPL one, change the region count, or enable or disable the replication feature, you use SSH to log into an HCP node. To do this, you need a laptop computer connected to the HCP system front- or backend network. For information on this, see Running HCP Setup from a service shell on page 3-9. This chapter describes the procedure for each of the activities listed above. For information on creating a user account with the service role, see Administering HCP.
92
Notes:
If the system is configured for DNS and you change the DNS name of
the system, make sure you change the name in the corporate DNS setup as well. For consistency, if you change the hostname prefix, you should also change that in the DNS setup.
After you change the DNS name, any clients connected to the System
Management Console at the time of the change need to close all browser windows to clear cookies so they can access the Console using the new name. This also applies to clients accessing the Tenant Management Console. Important: Do not change the DNS settings for an HCP system while any nodes are unavailable. If you do, those nodes will have incorrect DNS information when they rejoin the system. For more information on using DNS with HCP, see Domain name system on page 2-4. To change DNS settings for an HCP system: 1. In the top-level menu in the System Management Console, mouse over Configuration to display a secondary menu. 2. In the secondary menu, click on DNS Settings. 3. On the DNS Settings page:
Select or deselect the Use DNS option to have HCP use or not use DNS, respectively. To change the DNS name of the HCP system, in the System Name field, type the new DNS name for the system. This name can contain only lowercase letters, numbers, and hyphens (-). Additionally, it must consist of at least three segments, separated by periods (.).
93
To change the hostname prefix, in the Hostname Prefix field, type the new hostname prefix. The hostname prefix can be from one through 12 characters long and can contain only lowercase letters, numbers, and hyphens (-). Tip: To make node names easier to read, end the hostname prefix with a hyphen (-).
To change the DNS servers HCP uses, in the DNS Servers field, type a comma-separated list of the IP addresses of all the current corporate DNS servers. Spaces are allowed.
4. Click on the Update Settings button. A warning message appears asking you to confirm the changes youve made. 5. In the field in the message window, type YES. This is case sensitive. 6. Click on the Update Settings button. The HCP system restarts with the new settings. This takes a few minutes, during which the system is inaccessible to users and applications.
The IP address of the gateway for routing requests from the HCP
system to the front-end network
The IP address the HCP nodes use to multicast messages to all other
nodes
The maximum packet sizes for data sent on the front-end and back-end
networks
For each node, the front-end IP address and subnet mask and the backend IP address and subnet mask For an HCP SAIN system that uses CB 320 servers, the network configuration also includes the IP addresses of the server chassis. For information on modifying this list, see Adding and removing chassis IP addresses below.
94
This network configuration is set during the Appliance OS and HCP core software installation. Changes in the network topology, however, or a move to a new location may make changes to this configuration necessary. You can change the first three items listed above at any time while the HCP system is running. You can change the network settings for a node only while the node is running and available. All nodes must be on the same front-end subnet and the same back-end subnet. If you change the front- or back-end subnet for one node, you need to change it for all nodes. You should do this in a single operation. Important: Do not change the multicast IP address while any nodes are unavailable. If you do, the unavailable nodes will not be able to rejoin the system. Notes:
You cannot change the last octet in the back-end IP address for a node. You cannot change the back-end subnet mask for a node to anything
other than 255.255.255.0. For more information on the gateway router and the front- and back-end IP addresses and subnet masks for the HCP nodes, see Information for an OS installation on page 2-2. For more information on the multicast address for an HCP system, see Networking on page 2-4. To change network settings for an HCP system: 1. In the top-level menu in the System Management Console, mouse over Configuration to display a secondary menu. 2. In the secondary menu, click on Network Settings.
95
To change the IP address of the front-end gateway router, in the Default Gateway field, type the new gateway router IP address. Note: Changing this IP address may disrupt the network connection between your computer and the HCP system, thereby causing the System Management Console session to end. If this happens, start a new Console session on a computer with access to the new front-end network.
To change the multicast IP address for the HCP system, in the Multicast Address field, type the new multicast IP address. This address must begin with 238. To change the maximum packet size for the front-end network, in the Front-end MTU field, select the packet size you want. To change the maximum packet size for the back-end network, in the Back-end MTU field, select the packet size you want. To change network settings for an available node, do one or more of these:
In the Front-end IP field for the node, type a new front-end IP address. Note: If you change the front-end IP address for the node to which your System Management Console session is connected, the session ends. If this happens, start a new session.
In the Front-end Netmask field for the node, type a new subnet mask for the front-end IP address. In the Back-end IP field for the node, type a new back-end IP address. In the Back-end Netmask field for the node, type a new subnet mask for the back-end IP address.
4. Click on the Update Settings button. A warning message appears asking you to confirm the changes youve made.
96
5. In the field in the message window, type YES. This is case sensitive. 6. Click on the Update Settings button. The HCP system restarts with the new settings. This takes a few minutes, during which the system is inaccessible to users and applications.
To remove a single IP address from the list, click on the delete control ( ) for that address. To remove all the IP addresses from the list, click on Delete All.
97
When a shift occurs in the current system time, HCP does not adjust
object retention times. As a result of a shift forward, some objects due to expire in the future may immediately become expired. As a result of a shift backward, some objects may take longer to expire than originally intended. HCP stores all times (such as creation dates and retention settings) in Coordinated Universal Time (UTC) and uses its time zone setting only for display purposes. Changing the time zone for an HCP system, therefore, has no affect on object retention. For more information on HCP time settings, see Time on page 2-5. To change time settings for an HCP system: 1. In the top-level menu in the System Management Console, mouse over Configuration to display a secondary menu. 2. In the secondary menu, click on Time Settings.
98
To change the time source, in the Time Servers field, type either internal or the IP addresses or hostnames of one or more time servers. Use commas to separate multiple IP addresses and hostnames. To change the current system time when the time source is internal, in the Current Time field, type the new time. The format for the time is MMDDhhmmYYYY, where MM is the two-digit month, DD is the two-digit day, hh is hours on a 24-hour clock, mm is minutes, and YYYY is the year. If the time source is internal and you leave this field blank, the current system time doesnt change.
To change the time zone, in the Time Zone field, select the new time zone.
4. Click on the Update Settings button. A warning message appears asking you to confirm the changes youve made. 5. In the field in the message window, type YES. This is case sensitive. 6. Click on the Update Settings button. The HCP system restarts with the new settings. This takes a few minutes, during which the system is inaccessible to users and applications.
99
To change the serial number of an HCP system: 1. In the top-level menu in the System Management Console, mouse over Configuration to display a secondary menu. 2. In the secondary menu, click on Miscellaneous Settings. 3. In the Serial Number field, type the new serial number for the system. Valid values can contain only alphanumeric characters, hyphens (-), underscores (_), number signs (#), and spaces. 4. Click on the Update Settings button.
910
Zero-copy failover is first enabled or disabled during HCP installation. Its automatically disabled and then reenabled during an online upgrade of the HCP system. You can enable or disable zero-copy failover at any time while the HCP system is running. However, you should do this only when specifically told to do so by HCP support personnel. For more information on zero-copy failover, see Zero-copy failover on page 1-9. To enable or disable zero-copy failover: 1. In the top-level menu in the System Management Console, mouse over Configuration to display a secondary menu. 2. In the secondary menu, click on Miscellaneous Settings. 3. On the Miscellaneous Settings page, select or deselect the Enable Zerocopy Failover option, as applicable. 4. Click on the Update Settings button.
Enabling DPL one does not cause any change to either the system-level
DPL setting or any namespace DPLs.
With DPL one enabled, HCP system administrators can change the
system-level DPL setting to one, and tenant administrators can set namespace DPLs to one.
911
If DPL one is disabled while the system-level DPL setting is one, that
setting automatically changes to two.
If DPL one is disabled while the DPL for a namespace is explicitly set to
one (not dynamic), the namespace DPL does not change. To enable or disable DPL one: 1. Using the service SSH key, use SSH to log into any storage node in the HCP system as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 2. Do either of these:
912
To accommodate a growing or shrinking HCP system, you can explicitly change the number of regions (referred to as the region count). The minimum and maximum values for the region count are determined by the system configuration. The number you specify:
Must be at least four times the number of storage nodes in the system Cannot be less than one quarter of the current region count Cannot be greater four times the current region count
As a general rule, an HCP system should have from eight to twelve regions per storage node. When you change the region count, HCP splits or merges existing regions to arrive at the new number of regions. While this is happening, HCP services may pause temporarily. Additionally, system responses to client requests may be slightly slower. You cannot change the region count while the HCP system is being upgraded or while a data migration is in progress. Conversely, you cannot start an upgrade or data migration while the number of regions is changing. You can cancel a change to the region count while HCP is in the process of making the change. When you do this, HCP finishes any split or merge that was in progress, and the region count then remains where it is at that moment. The count is not rolled back to the previous number of regions.
913
The HCP 4.1 Configuration menu appears. (Option v is present only for SAIN systems.)
4. From the HCP 4.1 Configuration menu, enter s to display the HCP Service menu. 5. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. When you enter y or yes, the HCP Service menu appears.
914
6. From the HCP Service menu, enter 7 to display the Change Region Count menu.
7. Do either of these:
HCP Setup displays the current number of regions in the HCP system and prompts for the new region count.
2. Enter the number of regions you want. 3. In response to the confirming prompt, enter y or yes to confirm
your entry or n or no to try again. When you enter y or yes, HCP begins splitting or merging regions, as applicable, and the HCP Service menu reappears.
915
confirm your entry or n or no to try again. When you enter y or yes in both step 2 and this step, HCP stops splitting or merging regions, as applicable, and the HCP Service menu reappears. 8. From the HCP Service menu, enter q to return to the HCP 4.1 Configuration menu. 9. In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again. 10.From the HCP 4.1 Configuration menu, enter q to log out of the install shell. 11.In response to the confirming prompt, enter y or yes to confirm your entry or n or no to try again.
916
2. Enter:
admin jvm replication enable
system. This text can be up to 1,024 characters long and can contain any UTF-8 characters, including white space.
e. Click on the Restart System button.
917
918
10
Managing an HCP system
For a variety of reasons, a few HCP system management activities are considered to be more advanced than general system monitoring and configuration. These activities are:
Managing HCP services Resetting the checkpoint for a replication link Disabling chained writes Accessing the administrative interface for the HCP search facility
To manage HCP services, you need an HCP System Management Console user account with the service role. To reset replication checkpoints or disable chained writes, you use SSH to log into an HCP node. To do this, you need a laptop computer connected to the HCP system front- or back-end network. For information on this, see Running HCP Setup from a service shell on page 3-9. This chapter describes the procedure for each of the activities listed above. For information on creating a user account with the service role, see Administering HCP.
101
Service management
Normally, HCP services require no human intervention; they run according to a schedule or when triggered by specific events or are ongoing. With an HCP System Management Console user account with the monitor or administrator role, you can monitor services. However, with these roles, you have limited ability to affect when and how most services work. With the service role, you have the ability to start, stop, enable, and disable individual services:
Starting a service starts a full service run unless the service is already
running. If the service is already running, starting it has no effect.
Stopping a service stops the current run. Enabling a service allows it to run as it normally would. Enabling a
service does not start it.
102
You can perform all of these actions on all the services except indexing, shredding, and replication. Examples of when you would perform these actions include starting a particular service to force repairs ahead of schedule or disabling a service to temporarily lighten the load on the system. Services, except indexing, shredding, and replication, use internal storage segments to limit the load they put on the system. When running as scheduled or when triggered, each service processes one segment on each node during each run. When started manually, however, each service processes all the segments on each node sequentially in the same run. This can significantly increase the load the service puts on the system. The indexing service is ongoing and is managed differently from the other services. The shredding service is ongoing. You cannot start, stop, enable, or disable it. The replication service works through a separate mechanism. For more information on managing this service, see Replicating Tenants and Namespaces.
To start a service, click on the start control ( To stop a service, click on the stop control (
To enable a service, click on the enable control ( service. To disable a service, click on the disable control ( service.
) for that
103
In this output, the linkName field specifies the name of the replication link, and the linkId field specifies the link ID. Finding the link ID and resetting the checkpoint To find the ID of a replication link and reset the checkpoint for that link: 1. Using the service SSH key, use SSH to log into any storage node as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 2. Enter this command to list the currently configured replication links:
admin jvm replication list
104
3. In the list of replication links, find the ID for the link you want. 4. Enter this command to reset the checkpoint for the link:
replication setProgress link-id checkpoint-time
checkpoint-time is the time to which you want to reset the checkpoint, specified as seconds since January 1, 1970, at 00:00:00. For example, this command resets the checkpoint for link MA-CA to October 20, 2010, at 9:45 a.m.:
replication setProgress cac3d212-f2b1-4b42-9891-e8cb0a25372d 1287567900
105
3. Using a text editor such as vi, open the init.conf file for editing. 4. Remove the comment character (#) from this line:
# RISFS_DATA_CHAIN="--disable-chaining"
106
11
Troubleshooting an HCP system
The HCP system is designed to be self-healing and to run without human intervention. Occasionally, however, things do go wrong. As a service provider, you can do some basic troubleshooting, but more often, your role is to help HCP support personnel resolve customer problems. This chapter describes some actions you can take to help troubleshoot problems with the HCP system:
Understanding service alerts Retrieving the HCP Setup logs Resetting the install user password Reset a security user password Setting HCP internal logging levels Downloading the logs for a single node Creating a snapshot of the HCP JVM Using the HCP system service tools Viewing the HCP system status pages Ejecting the CD tray from a node
111
Messages
Online upgrade in progress
Description
The HCP system software is currently being upgraded while the system is online. For information on upgrading, see Chapter 4, Upgrading an HCP system.. The HCP system software has been upgraded with the system offline, and the upgrade now needs to be completed. The HCP system has been upgraded. You need to commit the upgrade now. For instructions on doing this, see Step 6: Complete the upgrade on page 4-25. For an HCP system that uses SAN-attached storage, the storage is configured for multipathing. For information on this feature, see Multipathing on page 1-9. Zero-copy failover is enabled. For information on this feature, see Zero-copy failover on page 1-9. Zero-copy failover is disabled.
Multipath enabled
ZCF enabled
ZCF disabled
All data and metadata in the repository is encrypted using the specified encryption algorithm. For information on this feature, see Security on page 2-8.
112
113
node-ip-address-list is a comma-separated list of the back-end IP addresses of all the nodes (both storage and search) in the HCP system. Spaces are not allowed.
For example:
sudo /sbin/install_passwd -c 10.4.1.15,10.4.1.16,10.4.1.17,10.4.1.18,10.4.1.19
3. Enter a new password for the install user. (As you type, the characters do not show.) The new password must follow the standard rules for Unix passwords. In particular, to be valid, a password must include at least one character from two of these three groups: alphabetic, numeric, and special characters. Also, a password cannot be a word found in the dictionary. A prompt to confirm the new password appears.
4. Enter the new password again. The install user password changes on all nodes in the HCP system.
114
To find a user ID, you use a command that lists all the currently defined user accounts. The output of this command includes the user ID, username, roles, and other information about each account. Heres an example of the output for one account:
'authenticationScheme' : 'LOCAL', 'description' : 'Primary HCP administrator', 'enableAtTime' : 'Wed Dec 31 19:00:00 EST 1969', 'failedLoginCount' : '2', 'fullName' : 'Lee Green', 'isEnabled' : True, 'isHidden' : False, 'isPasswordChangeRequired' : False, 'isPaswordExpired' : False, 'lastLoginTime' : 'Wed Nov 18 12:12:42 EST 2009', 'lastPasswordChangeTime' : 'Mon Nov 09 10:36:46 EST 2009', 'name' : 'lgreen', 'parentUUID' : '34142e35-6d9d-4d99-acd4-85656444a29a', 'roles' : 'MONITOR,SYSTEM,SECURITY', 'userId' : '101'
To find the ID of a security user and reset that users password: 1. Using the service SSH key, use SSH to log into any storage node in the system as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 2. Enter this command to list the currently defined user accounts:
admin jvm adminUser list
3. In the list of user accounts, find the user ID you want. 4. Enter this command to reset the password for the user you want:
admin jvm adminUser update cluster user-id --password new-password
For example:
admin jvm adminUser update cluster 101 --password Chang3Me!
115
Customers can use the HCP System Management Console to download the internal logs to send to the HCP support center. The logs can help support personnel identify and resolve problems with the system. A log download from the System Management Console includes all the logs for each node. As a service provider, you can also download the logs for a single node. When HCP prepares the internal logs for downloading on a node thats a Hitachi or Dell server, it includes the file in which it saves the server system event log. For information on this file, see Server system event log on page 11-8. For information on downloading the HCP internal logs from the System Management Console, see Administering HCP. For information on downloading the logs for a single node, see Downloading the logs for a single node on page 11-12.
116
For the Java virtual machine (JVM), which has multiple components,
messages are written to different files based on their severity level. Also, HCP creates multiple instances of the JVM log files to support concurrent logging. Each file name has this format:
jvm1.node-number.instance.severity-level
For example, one of the current files with WARNING-level internal messages for the JVM component on the node with the back-end IP address 10.4.1.15 is jvm1.15.0.WARNING. For both the C and Java components, the name of the current file containing the standard output and error messages has this format:
R00.node-number.component-name-out
For example, the current file with standard output and error messages for the timelord component on the same node as above is R00.15.timelord-out. Daily logs During any given day, if the .log file for a component grows larger than three MB, HCP appends .0 to the file name and starts a new .log file. For example, if the .log file for the timelord component on node 10.4.1.15 grows larger than three MB, HCP renames the file timelord.15.log.0 and creates a new timelord.15.log file. If it happens again, HCP increments the timelord.15.log.0 file name to timelord.15.log.1, changes the name of the timelord.15.log file to timelord.15.log.0, and creates a new timelord.15.log file. This can repeat up to 99 times, after which HCP deletes the oldest file each time a new file is started. If, during the day, a node restarts, HCP increments the first part of the -out file name for each component and creates a new -out file. For example, if node 10.4.1.15 restarts, HCP changes the name of the R00.15.timelord-out file to R01.15.timelord-out and creates a new R00.15.timelord-out file. This can also repeat up to 99 times. Log rotation HCP rotates all the log files on a node daily at 4:02 a.m. It does this by compressing all the files into a single tar file with this name:
logs-YYYYMMDD-hhmmss.tar.bz2
117
It stores this file in the /var/ris directory. If the storage allocated for the /var/ris directory fills beyond 70 percent, HCP moves all the tar files in the directory to the /RIS/archivehighest-archive-number directory. If that directory fills beyond 70 percent, HCP deletes the tar files in it one at a time, starting with the oldest, until the directory storage is less than 70 percent full. Each time it rotates the log files, HCP checks both the /var/ris and /ris/archivehighest-archive-number directories for tar files that are over 35 days old. If it finds any, it deletes them.
Logging levels
Logging levels correspond to message severity levels:
118
FINE Messages at this level record basic sequence execution. INFO Messages at this level record normal behavior. This is the
default logging level for C components.
In this format:
node-ip-address is the back-end IP address of the node on which you want to change the logging level.
119
port-number is the component-dependent port number. The table below shows the port number for each component for which you can set the logging level.
Component
mailgate risfs-0 risfs-1 slabd storman-n 2500 20000 20001 1234 10000 + n Each logical volume has its own STORMAN process. These processes are numbered starting from 1 (one) on each node. sysmon timelord 12000 12001
Port
logging-level is one of the logging levels listed above. This is case sensitive.
3. Use a text editor such as vi to open the log.properties file for editing.
1110
4. Add a line specifying the new logging level to the end of the file. For this line, use the exact format provided by HCP support personnel. Heres an example of a line that specifies a new logging level for the com.ris.server component of the JVM:
com.ris.server.level=INFO
To change the logging levels of multiple components, add one line for each change. 5. Save and close the log.properties file. 6. To put the new logging level into effect, enter:
/opt/ris/bin/admin RELOAD_JAVA_LOG_CONFIG
Tip: To make the same change on all other nodes: 1. Use the cluster_put tool to propagate the log.properties file to the other nodes. 2. Use the cluster_ssh tool to execute the admin command above on the other nodes. For information on cluster_put and cluster_ssh, see Using the system service tools on page 11-15. Changing JVM components back to the default logging level To change the logging levels for all JVM subcomponents on a node back to the default: 1. Using the root SSH key, use SSH to log into the node as root. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 2. Change to the /opt/ris/config directory:
cd /opt/ris/config
3. Copy the existing log.properties file to log.properties.n, where n creates a unique name for the file; for example:
cp log.properties log.properties.1
1111
This saves the file with the nondefault logging levels, so you can use it again if needed. 4. Copy the /opt/ris/config/log.properties.nologs file to /opt/ris/config/log.properties:
cp log.properties.nologs log.properties
In this format:
newest-day is the number of days ago for the most recent logs you want. number-of-days is the number of days of logs to download, working backwards in time from the value specified for newest-day.
nocrypt specifies that the downloaded logs should not be encrypted.
output-file-spec.tar.bz2.gpg specifies the directory path and file name for the compressed file containing the downloaded logs. output-filespec can include an absolute path or a path relative to the current directory.
1112
3. Using an SCP tool such as scp or winscp, download the file with the collected logs to another computer. Tip: scp is already installed on each node in the HCP system.
This command collects only the logs from the current day:
sudo /opt/arc/bin/logdl 0 0 > temp/todaysLogs.tar.bz2.gpg
This command collects the logs from seven days ago through six days
ago (that is, it packages two days worth of logs):
sudo /opt/arc/bin/logdl 6 2 > temp/oldLogs.tar.bz2.gpg
Heres an example of an scp command that downloads the file created by the first logdl command shown above:
scp temp/todaysLogs.tar.bz2.gpg lgreen@192.168.57.42:/home/lgreen/hcpLogs
1113
In addition to uploading an encryption key, you can list the keys HCP currently has and delete any of those keys. When you list the keys, the information displayed for each key includes the key fingerprint. When deleting a key, you identify it by its fingerprint. Important: Do not delete the system-supplied public encryption key or signing sub key, which are identified by <support@hds.com> in the user ID. If you do, support personnel will not be able to decrypt the downloaded logs. To upload, list, and delete the encryption keys used for log downloads: 1. Using the service SSH key, use SSH to log into any storage node in the HCP system as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 2. Do one of these:
For example:
admin jvm logdl addKey /home/service/temp/logEncryptKey
For example:
admin jvm logdl deleteKey E96FFEA44F67290FBF2DF49D77D9ED35CD2B3BF8
1114
When you generate the JVM snapshot, HCP appends it to a file in the /var/ ris/logs directory, which is the directory that contains the HCP internal logs. The name of this file is always R00.node-number.jvm-out (for example, R00.16.jvm-out). At any time, if the JVM on a node restarts, HCP increments the first part of the file name to R01 and creates a new R00.node-number.jvm-out file. If the JVM restarts again during the same day, HCP increments R01 to R02 and R00 to R01. This can repeat up to 99 times. When HCP rotates the log files each day, it includes any JVM snapshot files that start with R01 or higher. For information on how HCP rotates the log files, see Log rotation on page 11-7. When the HCP internal logs are downloaded either from the System Management Console or from an individual node, the JVM snapshot files for the applicable time period are automatically downloaded with them. To generate a snapshot of the JVM on a node: 1. Using the service SSH key, use SSH to log into the node as service. For instructions on doing this, see Appendix D, Setting up SSH key authentication. 2. To generate the snapshot files, enter:
sudo /opt/arc/bin/generate-thread-dump.sh.
1115
You can use the -n or --node parameter in the tool command to specify
the front-end IP address of a node in the target system.
You can set the TARGET_CLUSTER environment variable to the frontend IP address of a node in the target system. The tool looks first for an -n, --node, -c, or --config-file parameter. If none is present, it looks for the TARGET_CLUSTER variable. If this variable is not set, it uses the current system. Target nodes By default, the system service tools act on all nodes in the target HCP system. However, you can limit the requested operation to only one type of node:
You can use the -a parameter in the tool command to limit the
operation to only storage nodes.
1116
You can use the -s parameter in the tool command to limit the
operation to only search nodes. If you specify both the -a and -s parameters in a single command, the tool returns an error. System service tool output Each system service tool writes all output both to the console and to log files on the node from which you run the tool. The log files, one for each target node, are located in the /opt/arc/tools/cluster/logs directory. The name of each log file is node-ip-address.output (for example, 192.168.210.16.output). The system service tools all use the same set of log files. As each tool runs, it appends its output to the text that already exists in the files.
cluster_ssh
The cluster_ssh tool does any of these:
Reboots or shuts down each target node Retrieves internal log messages from the target nodes
Syntax for this tool is:
cluster_ssh.py (-h|--help) |--version |([(-c|--config-file) config-file-spec] [(-n|--node) node-ip-address] [-a|-s] [(-u|--user) username] [-v|--verbose] (((-l|--level) message-severity-level) |(-r|--reboot) |(-H|--halt) |((-i|--inputScript) script-file-spec) |"command-string"))
1117
cluster_ssh parameter descriptions The table below describes the parameters for the cluster_ssh command.
Parameter -h|--help --version (-c|--config-file) config-file-spec Description
Displays and describes the syntax for the cluster_ssh command; then exits. Displays the version number of the currently installed cluster_ssh tool; then exits. Specifies an HCP configuration file that identifies the system on which cluster_ssh should act. The configuration file you specify must be located on the node you logged into and must identify the system that includes that node. For information on using this parameter, see Target HCP system on page 11-16.
(-n|--node) node-ip-address
Specifies a node that identifies the HCP system on which cluster_ssh should act. node-ip-address is the front-end IP address of the node. For information on using this parameter, see Target HCP system on page 11-16.
Tells cluster_ssh to act only on storage nodes. Tells cluster_ssh to act only on search nodes. Tells cluster_ssh to log into each node in the target system with the specified username. Reports cluster_ssh activity in greater detail. If an error occurs while cluster_ssh is running, rerunning the same command with this parameter can help identify the problem. Retrieves all the HCP internal log messages with the specified severity level or a more severe level from the log files in the /var/ris/logs directory on each target node. Valid values for message-severity-level are: FINEST FINER FINE INFO WARNING SEVERE These values are case sensitive. This parameter is mutually exclusive with the -r, --reboot, -H, --halt, -i, --inputScript, and command-string parameters. For information on the HCP internal logs, see Managing the HCP internal logs on page 11-5.
(-l|--level) messageseverity-level
1118
Parameter -r|--reboot
Description
Tells cluster_ssh to shut down HCP on each target node and then reboot the node. This parameter is mutually exclusive with the -l, --level, -H, --halt, i, --inputScript, and command-string parameters.
-H|--halt
Tells cluster_ssh to shut down HCP on each target node and then power off the node. This parameter is mutually exclusive with the -l, --level, -r, --reboot, -i, --inputScript, and command-string parameters. Note: If any node does not power off automatically, wait 30 minutes before powering it off manually. If the node is connected to a console, the console displays System halted when you can safely power off the node.
(-i|--inputScript) script-file-spec
Specifies a user-supplied script for cluster_ssh to execute on each target node. script-file-spec is the name of the script file with or without a directory path. If included, the path can be absolute or relative to the current directory. After running the specified script, cluster_ssh removes the script from the node. This parameter is mutually exclusive with the -l, --level, -r, --reboot, -H, --halt, and command-string parameters.
1119
Parameter "command-string"
Description
Specifies one or more commands for cluster_ssh to execute on each target node. cluster_ssh executes multiple commands on any given node in the order in which you specify them. Use a semicolon (;) or a double ampersand (&&) to separate each command from the next: The semicolon means proceed with the next command even if the current command fails. The double ampersand means do not proceed if the current command fails.
The double quotation marks (") are required when you specify more than one command or when you specify a single command with one or more options. You can omit the quotation marks around a single command with no options. This parameter is mutually exclusive with the -l, --level, -r, --reboot, -H, --halt, -i, and --inputScript parameters. Note: To reboot nodes, use the -r or --reboot parameter instead of a reboot command. With a reboot command, the current node may shut down before it can communicate the request to the other target nodes. Similarly, use the -H or --halt parameter instead of a shutdown command.
1120
1121
1122
1123
Using the system service tools ================== || 10.4.1.17: ================== /var/ris/logs/R01.17.storman1-out:10/10 11:28:11 storman-1 20713 17 (SEVERE): PANIC AT register.c:93:restart_admin Unable to connect to the admin after 2 tries /var/ris/logs/R01.17.storman2-out:10/10 11:28:12 storman-2 20749 17 (SEVERE): PANIC AT register.c:93:restart_admin Unable to connect to the admin after 2 tries . . . /var/ris/logs/sysmon.17.log:10/10 11:18:01 sysmon 3418 17 (SEVERE): Admin handler returned failure: -1 Broken pipe /var/ris/logs/sysmon.17.log:10/10 11:18:03 sysmon 3418 17 (SEVERE): PANIC AT register.c:93:restart_admin Unable to connect to the admin after 0 tries ==================== 3 Success 0 Error ==================== Command results saved at /var/opt/arc/tools/cluster/logs
cluster_get
The cluster_get tool retrieves one or more files from each target node and stores them on the current node. By default, the files are stored in subdirectories, one per node, of a specified directory, but you can choose to have them stored directly in that directory. The name of each subdirectory is the IP address of the applicable node. If the cluster_get command uses -n or --node parameter to identify the target system, this is the front-end IP address. Otherwise, its the back-end IP address. Syntax for the cluster_get tool is:
cluster_get.py (-h|--help) |--version |([(-c|--config-file) config-file-spec] [(-n|--node) node-ip-address] [-a|-s] [(-u|--user) username] [-d|--disable-unique-ify] [-v|--verbose] ((-l|--logs) |(-i|--image) |((source-file-spec|source-path) destination-path))
1124
cluster_get parameter descriptions The table below describes the parameters for the cluster_get command.
Parameter -h|--help --version (-c|--config-file) config-file-spec Description
Displays and describes the syntax for the cluster_get command; then exits. Displays the version number of the currently installed cluster_get tool; then exits. Specifies an HCP configuration file that identifies the system on which cluster_get should act. The configuration file you specify must be located on the node you logged into and must identify the system that includes that node. For information on using this parameter, see Target HCP system on page 11-16.
(-n|--node) node-ip-address
Specifies a node that identifies the HCP system on which cluster_get should act. node-ip-address is the front-end IP address of the node. For information on using this parameter, see Target HCP system on page 11-16.
Tells cluster_get to retrieve the requested files only from storage nodes. Tells cluster_get to retrieve the requested files only from only on search nodes. Tells cluster_get to log into each target node with the specified username. Puts all the returned files directly in the directory identified by the destination-path parameter. With this parameter, files with the same name that come from different nodes overwrite each other Without this parameter, cluster_get puts each returned file in a directory named with the IP address of the node from which it was retrieved. These directories are located in the directory specified by the destination-path parameter. This parameter is not valid with the -l, --logs, -i, and --image parameters.
-v|--verbose
Reports cluster_get activity in greater detail. If an error occurs while cluster_get is running, rerunning the same command with this parameter can help identify the problem.
1125
Parameter -l|--logs
Description
Retrieves and compresses the HCP internal log files from the /var/ris/logs directory on each target node. The log files from each node are compressed into a single file named node-front-end-ip-address_logs.tar.gz, which is then stored in a node-specific subdirectory of the current directory. The name of each subdirectory is the IP address of the applicable node. If the cluster_get command uses -n or --node parameter to identify the target system, this is the front-end IP address. Otherwise, its the backend IP address. If you use this parameter, do not include the -i, --image, source-filespec, source-path, or destination-path parameter in the cluster_get command. For information on the HCP internal log files, see Managing the HCP internal logs on page 11-5.
-i|--image
Retrieves and compresses the HCP internal log files from the /var/ris/logs directory on each target node, along with other node information thats useful for identifying problems. The log files and other information from each node are compressed into a single file named node-front-end-ipaddress_image.tgz, which is then stored in a node-specific subdirectory of the current directory. The name of each subdirectory is the IP address of the applicable node. If the cluster_get command uses -n or --node parameter to identify the target system, this is the front-end IP address. Otherwise, its the backend IP address. If you use this parameter, do not include the -l, --logs, source-filespec, source-path, or destination-path parameter in the cluster_get command. For information on the HCP internal log files, see Managing the HCP internal logs on page 11-5. Important: This operation is CPU intensive. It also makes it more difficult for support personnel to review the logs in real time. Do not do this unless explicitly told to do so by HCP support personnel.
source-file-spec
Specifies the file cluster_get should retrieve from all the target nodes. The file specification must include the absolute path to the file.
1126
Parameter source-path
Description
Specifies a directory from which cluster_get should retrieve all files, including the files in any directories below that directory (that is, all subdirectories, recursively). cluster_get retrieves the files from all the target nodes.
The cluster_get command below retrieves the build.version file from each
HCP node and stores it in a directory named with the back-end IP address of that node. cluster_get creates these directories in the current directory. You can use this command to verify that the same version of HCP is installed on all nodes in the system.
cluster_get.py /opt/arc/build.version ./
1127
1128
cluster_put
The cluster_put tool stores one or more files on each target node. Syntax for this tool is:
cluster_put.py (-h|--help) |--version |([(-c|--config-file) config-file-spec] [(-n|--node) node-ip-address] [-a|-s] [(-u|--user) username] [-v|--verbose] (source-file-spec|source-path)) [destination-path])
1129
cluster_put parameter descriptions The table below describes the parameters for the cluster_put command.
Parameter -h|--help --version (-c|--config-file) config-file-spec Description
Displays and describes the syntax for the cluster_put command; then exits. Displays the version number of the currently installed cluster_put tool; then exits. Specifies an HCP configuration file that identifies the system on which cluster_put should act. The configuration file you specify must be located on the node you logged into and must identify the system that includes that node. For information on using this parameter, see Target HCP system on page 11-16.
(-n|--node) node-ip-address
Specifies a node that identifies the HCP system on which cluster_put should act. node-ip-address is the front-end IP address of the node. For information on using this parameter, see Target HCP system on page 11-16.
Tells cluster_put to store the specified files only on storage nodes. Tells cluster_put to store the specified files only on search nodes. Tells cluster_put to log into each target node with the specified username. Reports cluster_put activity in greater detail. If an error occurs while cluster_get is running, rerunning the same command with this parameter can help identify the problem. Specifies the file cluster_put should store. cluster_put takes this file only from the node on which you issue the command. The file specification can include an absolute path or a path relative to the current directory.
source-file-spec
source-path
Specifies a directory containing the files cluster_put should store, including the files in any directories below that directory (that is, all subdirectories, recursively). cluster_put takes these files only from the node on which you issue the command.
1130
Parameter destination-path
Description
Specifies the directory in which cluster_put stores the files on each target node. The specified directory must already exist.
Tip: To create the /opt/arc/tools/cluster/temp directory on all the nodes in the system, you could use the cluster_ssh tool.
1131
The cluster_put command below copies the temp/cluster.cfg file from the
current node to all the nodes in the HCP system, overwriting the cluster.cfg file in the /opt/arc/config directory on each. The command tells cluster_put to log into each node with the username root in order to have the permissions required for the requested action. This command is useful for propagating configuration changes to all the nodes in the system.
cluster_put.py -u root temp/cluster.cfg /opt/arc/config
Important: Do not modify the cluster.cfg file unless explicitly told to do so by HCP support personnel. The console output from this command looks like this:
NODES: Storage: 10.4.1.15 10.4.1.16 Search: 10.4.1.17 Copying local file temp/cluster.cfg to [NODES]:/opt/arc/config ==================== 3 Success 0 Error ====================
1132
For example:
https://admin.hcp-ma.example.com:8000/cluster/getAdminView.action
Note: To view the system status pages, you must be in an active HCP System Management Console session. If your session times out due to inactivity, you cannot display these pages.
1133
To eject the CD tray from an HCP node: 1. In the top-level menu in the HCP System Management Console, click on Nodes. 2. In the list of nodes, click on the number of the node you want. 3. At the bottom of the node details page, click on the Eject CD button.
1134
A
Forms
This appendix contains forms that you can use to facilitate some of the procedures presented in this book:
Encryption Key lets you record the encryption key displayed during HCP
installation processing.
HCP Node Addition Information lets you record the information you
need for a node addition.
A1
__________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ Front-end subnet mask: ______________________________________ Front-end network gateway IP address: _________________________
A2
________________________________________
DNS name of the HCP system: _________________________________ If DNS enabled, IP addresses of corporate DNS servers: ____________
__________________________________________________________ Time zone: Internals Storage configuration: If external, shared storage: If external, WMS-100 or AMS-200: If external, blade servers If blade servers, chassis IP addresses: Internal External Yes Yes Yes No No No ________________________________________________
__________________________
A3
Security Enable encryption: Storage nodes Back-end IP addresses of storage nodes (list, range, or file name): ____ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ Search nodes Back-end IP addresses of search nodes (list, range, or file name): _____ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ Yes No
A4
Encryption Key
Encryption Key
Use this form to record the key displayed during installation processing for an HCP system with encryption enabled. HCP system serial number: Encryption key: ___________________________________
____________________________________________
Caution: KEEP THIS KEY IN A SECURE LOCATION. Loss of this key will most likely result in unrecoverable data in the case of catastrophic system failure.
A5
install user password for logging into the HCP nodes: _______________ Version number of the Appliance Operating System currently installed on the existing HCP nodes: ___________________________________
Version number of the HCP software currently installed on the existing HCP nodes: ________________________________________________ Front-end IP address/back-end IP address of each new storage node: __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ Front-end IP address/back-end IP address of each new search node: __________________________________________________________ __________________________________________________________ __________________________________________________________ __________________________________________________________ Front-end subnet mask already set for the existing nodes in the system: __________________________________________________________ Front-end network gateway IP address already set for the existing nodes in the system: _______________________________________________
For SAIN systems that use CB 320 servers, chassis IP address of each new server: ___________________________________________________
A6
B
HCP Setup wizard menus
The program you use to install, upgrade, or add nodes to an HCP system is called HCP Setup. It uses a menu-based wizard to step you through the procedure you choose to perform. This appendix shows each of the wizard menus. It presents them in the order in which they occur when you follow the default path through the wizard for the applicable procedure. For general instructions on using the HCP Setup wizard, see HCP Setup program on page 3-7.
B1
Installation menus
Installation menus
Option 2 on the HCP 4.1 Configuration menu displays the wizard New Install menu.
When you select option 2 on the New Install menu, the wizard presents the HCP Networking Options menu.
When you select option 3 on the New Install menu, the wizard presents the HCP DNS Options menu.
B2
Installation menus
When you select option 4 on the New Install menu, the wizard presents the HCP Time Options menu.
When you select option 5 on the New Install menu, the wizard presents the Internal Configuration Settings menu.
When you select option 6 on the New Install menu, the wizard presents the HCP Security Settings menu.
B3
Upgrade menu
When you select option 7 on the New Install menu, the wizard presents the HCP Nodes menu.
Upgrade menu
Option 3 on the HCP 4.1 Configuration menu displays the Upgrade menu.
B4
B5
B6
C
HCP configuration file
The HCP Setup program you use to install the HCP software uses configuration information stored in a file named cluster.cfg. This file is located in the /home/install/arc-deploy directory on the node from which youre installing. Caution: Do not modify the cluster.cfg file in the /home/install/arc-deploy directory without explicit instructions from HCP support personnel. Doing so can cause the installation, upgrade, or node addition to fail or the HCP system not to run correctly after an installation, upgrade, or node addition. Once an HCP system is installed, you can view its cluster.cfg file in the /opt/ arc/config directory on each node. HCP Setup uses this file for system upgrades and node additions. Caution: Do not modify the cluster.cfg file in the /opt/arc/config directory. Doing so can cause upgrades and node additions to fail or the HCP system not to run correctly after an upgrade or node addition. This appendix contains a table of the parameters in the cluster.cfg file. The parameters are broken out by the sections of the file in which they occur. Within each section, they are presented in alphabetical order. For each parameter, the table provides a description and an example and indicates whether the HCP Setup wizard prompts for a value for it. For information on modifying the cluster.cfg file in the /home/install/arc-deploy directory, see Using a modified HCP configuration file on page 3-32.
C1
Parameter
install config_key
Description
Prompt
Version number of the configuration file. Example: config_key = 4.1.0 Caution! Do not change the value set by the HCP Setup wizard. Doing so will cause the installation or upgrade to fail.
No
type
Type of installation. Example: type = install-fresh Caution! Do not change the value set by the HCP Setup wizard. Doing so will cause the installation or upgrade to fail.
No
arc_nodes In /home/install/arc-deploy/ Back-end IP address of a storage node in the HCP system. The configuration file includes one line for each storage node. cluster.cfg: nodenode-number In the cluster.cfg file in the /home/install/arc-deploy directory, the node numbers in the parameter keywords start from 0 In /opt/arc/config/ (zero). cluster.cfg: node_node-IP-address In the cluster.cfg file in the /opt/arc/config directory, the IP address in each parameter keyword is the same IP address as the parameter value. Examples: In /home/install/arc-deploy/cluster.cfg: Yes
= = = =
In /opt/arc/config/cluster.cfg:
= = = =
C2
(Continued)
Parameter
search_nodes
Description
Prompt
In /home/install/arc-deploy/ Back-end IP address of a search node in the HCP system. The configuration file includes one line for each search node. cluster.cfg: nodenode-number In /opt/arc/config/ cluster.cfg: node_node-IP-address In the cluster.cfg file in the /home/install/arc-deploy directory, the node numbers in the parameter keywords start from 0 (zero). In the cluster.cfg file in the /opt/arc/config directory, the IP address in each parameter keyword is the same IP address as the parameter value. Examples: In /home/install/arc-deploy/cluster.cfg:
Yes
C3
(Continued)
Parameter
chassis_ip_csl
Description
Chassis IP addresses of the CB 320 servers used by the HCP system, if any. Use commas to separate multiple IP addresses. Spaces are not allowed. This parameter applies only when the blade_servers parameter is set to True. Example: chassis_ip_csl = 192.168.73.16
Prompt
Yes
clname
DNS name of the HCP system, specified as hcp-name. domain-name. Example: clname = hcp-ma.example.com
Yes
contact_info
Contact information for the HCP support center. The default depends on the distributor key. Example: contact_info = 'United States: (800) 446-0744.
Yes
current_time
Time on a 24-hour clock to which the HCP nodes initially set their system clocks, in this format:
Yes
hhmmssMMDDYYYY
This parameter applies only when the timeserver parameter is set to internal. Example: current_time = 13450003222011 currentmsgversion Used for upgrades only, version of the internal messaging protocol the existing HCP system uses. The HCP Setup wizard automatically supplies this value. Example: currentmsgversion = 410 disk_flush Option to have HCP flush the disk cache after every write operation. Valid values are True (flush the cache) and False (do not flush the cache). The default is False. Example: disk_flush = False distributor_key Name that determines the SSH keys required to use SSH to log into HCP nodes when the lockdown parameter is set to True. These SSH keys are initially provided by HDS and are unique for each authorized distributor of HCP. Yes No No
C4
(Continued)
Parameter
dns_enabled
Description
Option to use DNS with the HCP system. Valid values are True (use DNS) and False (do not use DNS). The default is true. If this parameter is set to True, the dns_servers parameter must specify the IP addresses of one or more DNS servers. Example: dns_enabled = True
Prompt
Yes
dns_servers
IP addresses of one or more corporate DNS servers. Use commas to separate multiple IP addresses. The default is the single IP address 192.168.100.10. For the HCP system to use DNS, you need to specify True for the dns_enabled parameter. Example: dns_servers = 192.168.107.10, 192.168.107.34
Yes
enable_replication
Option to allow the HCP system to replicate to another system or to be the target of replication. Valid values are True (allow replication) and False (disallow replication). The default is False. Example: enable_replication = True Note: Allowing replication doesnt cause the system to replicate. Replication starts with the creation of a replication link.
Yes
encryption
Option to encrypt data and metadata in the repository. Valid values are True (encrypt) and False (do not encrypt). The default is False. If this parameter is set to True, the encryption_cipher, encryption_ivmode, and encryption_key_length parameters must specify values. Example: encryption = True
Yes
encryption_cipher
Method used to encrypt the data and metadata in the repository. Valid values are the names of encryption methods. The default is AES. This parameter applies only when the encryption parameter is set to True. Example: encryption_cipher = AES
No
C5
(Continued)
Parameter
encryption_ivmode
Description
Initialization vector for the encryption cipher. This parameter applies only when the encryption parameter is set to True. Example: encryption_ivmode = essiv:sha256
Prompt
No
encryption_key_length
Length of the encryption key. Valid values are positive integers. This parameter applies only when the encryption parameter is set to True. Example: encryption_key_length = 256
No
filesystem
File system to use for storing objects. The only valid value is ext3. Example: filesystem = ext3
No
frontend_mtu
Maximum packet size, in bytes, for data sent on the front-end network. Valid values are positive integers. The default is 1500. Example: frontend_mtu = 1500
No
gateway
IP address of the front-end network gateway. The default is the gateway IP address specified during the Appliance OS installation. Example: gateway = 192.168.210.1
Yes
hostname_prefix
Prefix used in the names of nodes in the HCP system. This can be blank or a text string followed by a hyphen. Valid characters are lowercase letters, numbers, and hyphens (-). The complete hostname is formed by combining the prefix with the fourth octet of the back-end IP address for the node. The default for this parameter is hcp-node-. Example: hostname_prefix = hcp-node-
Yes
kerberos_ssh
Option to enable Kerberized SSH login to the HCP nodes. Kerberized SSH can be used for certain rare support situations. Valid values are True (enable Kerberized SSH login) and False (disable Kerberized SSH login). The default is False. Example: kerberos_ssh = False Note: Additional setup is required for Kerberized SSH to work. For information on this setup, see http:// www.visolve.com/security/ssh_kerberos.php.
No
C6
(Continued)
Parameter
lockdown
Description
Option to disable the ability to use SSH to log into HCP nodes without the SSH key supplied by the HCP distributor (applies after HCP Setup has finished processing). Valid values are True (disallow SSH without supplied key) and False (allow SSH without supplied key). The default is True. If this parameter is set to True, the distributor_key parameter must specify the name for the SSH keys provided by HDS. Example: lockdown = True Note: For customer installations, this parameter must be set to True.
Prompt
Yes
max_simul_mkfs
Maximum number of logical volumes HCP Setup will format at the same time on each node. Valid values are integers. The default is 4. The recommended maximum is 6. Example: max_simul_disks = 4
No
max_simul_nodes
Maximum number of nodes on which HCP Setup will simultaneously install HCP. Valid values are integers. The default is 10. However, if the HCP system is sharing external storage with other applications, this parameter is set to 2. Example: max_simul_nodes = 10
mcast
IP address of the multicast network for the HCP system. The default is 238.177.1.1. If two systems share back-end switches, each system must have a different multicast network and be on a different backend subnet. Example: mcast = 238.144.60.2
Yes
mdpl
Number of copies of the metadata for each object HCP must maintain in the repository. Valid values are 2, 3, and 4. The default is 2. Example: mdpl = 2 Caution! Do not set mdpl to 1 (one). Doing so can result in many irreparable objects if a node fails. Also, values higher than 2 have not been tested.
No
mm_filesystem
File system to use for the internal drives in CR 220 servers in HCP SAIN systems. The only valid value is ext3. Example: mm_filesystem = ext3
No
C7
(Continued)
Parameter
old_software_version
Description
Used for upgrades only, currently installed version of the HCP software. The HCP Setup wizard automatically supplies this value. Example: old_software_version = 4.0.1.169 Caution! Do not change the value set by the HCP Setup wizard. Doing so will cause the upgrade to fail.
Prompt
No
policy_enabled
Option to have services initially enabled on the HCP system. Valid values are True (services are enabled) and False (services are disabled). The default is True. This parameter applies to the capacity balancing, compression, content verification, disposition, duplicate elimination, garbage collection, protection, and scavenging services. Example: policy_enabled = True
No
regionmapsize
Number of regions to create. Valid values are integers greater than or equal to four times the number of storage nodes in the system. The default for this parameter is based on the initial number of storage nodes in the system. For a small system (for example, four nodes), the default is 32. Example: regionmapsize = 44
No
serial_number
Yes
single_node
Indication of whether the HCP system has one or more storage nodes. Valid values are True (the system has a single storage node) and False (the system has multiple storage nodes). The default is False. Example: single_node = False Note: Single-node systems are intended for testing purposes only. For customer installations, this parameter must be set to False. If this parameter is set to True and the cluster.cfg file lists multiple storage nodes, the installation or upgrade will fail.
No
C8
(Continued)
Parameter
software_version
Description
Version of the HCP software being installed. Example: software_version = 4.1.0.94 Caution! Do not change the value set by the HCP Setup wizard. Doing so will cause the installation or upgrade to fail.
Prompt
No
storage_type
Type of storage used by the HCP system. Valid values are internal (non-SAN-attached storage) and external (SANattached storage). The default is Not Set. Example: storage_type = external
Yes
timeserver
Source the HCP nodes use to adjust their system clocks. Valid values are the IP addresses of one or more servers outside the HCP system and internal, which causes the nodes to synchronize their system clocks to each other. Use commas to separate multiple IP addresses. The default is internal. Example: timeserver = internal
Yes
timezone
Time zone in which the HCP system is installed. Select the time zone from the list presented by the HCP Setup wizard. The default depends on the distributor key. Example: timezone = America/Los_Angeles
Yes
zero_copy_failover
Option to enable zero-copy failover. Valid values are true (enable zero-copy failover) and false (disable zero-copy failover). The default depends on the distributor key. Example: zero_copy_failover = true
Yes
C9
C10
D
Setting up SSH key authentication
To perform service procedures on HCP nodes, you need to use SSH to log into the applicable node. To use SSH, you need an SSH key. HCP has three SSH keys:
One for logging into the nodes before the HCP core software is installed
(the installation key)
One for logging into the HCP nodes as the service user after theyre
locked down (the service key)
One for logging into the HCP nodes as the root user after theyre locked
down (the root key) You get these keys, along with their passphrases, from the HCP distributor. The procedure for getting these keys depends on the distributor. To use SSH keys for HCP procedures, you first add the applicable keys to your keychain on your computer. Then you forward them to the node youre logging into. Each time you log into a node, you need to forward the applicable SSH keys. The procedure for adding and forwarding SSH keys differs for Windows and Linux. This appendix contains instructions for both. Caution: Never give the any of the HCP SSH keys to customers. Doing so puts the integrity of the stored data at risk.
D1
This file installs both PuTTY and Pageant. To add one or more SSH keys to your keychain on a Windows computer and forward them to an HCP node: 1. Start Pageant. 2. In the Pageant Key List window, click on the Add Key button. 3. In the Select Private Key File window, select the file containing the key you want. Then click on the Open button. 4. In the Pageant: Enter Passphrase window, type the passphrase for the key youre adding. Then click on the OK button. 5. Repeat steps 2 through 4 for any additional keys you need for the procedure you want to perform. 6. In the Pageant Key List window, click on the Close button. 7. Start PuTTY. 8. In the PuTTY Configuration window:
a. In the Category list, select Session (if its not the current category). b. In the Host Name (or IP address) field, type the IP address of the
D2
A PuTTY window opens and prompts for a username. Note: If a message window appears before the PuTTY window opens, click on the OK or Yes button to close it. 9. In response to the login prompt in the PuTTY window, enter service or root, as applicable, for the procedure you want to perform. Tip: If youre prompted for a password at this point, it means your SSH keychain does not include the correct SSH keys for what youre trying to do.
key:
chmod 600 /directory-path/key-file-name b. Add the key to your keychain: ssh-add /directory-path/key-file-name c. When prompted, enter the passphrase for the key.
D3
4. Use SSH with the -A option to log into the node you want as service or root, as applicable, for the procedure you want to perform:
ssh -A username@node-ip-address
Tip: If youre prompted for a password at this point, it means your SSH keychain does not include the correct keys for what youre trying to do.
D4
E
Accessing HCP interfaces through the back-end network
When diagnosing or resolving a problem with an HCP system, you may need to access one of these interfaces:
The HCP Configuration menu on a node in the system The HCP System Management Console The Tenant Management Console for a particular tenant The Namespace Browser for a particular namespace
HCP users access these interfaces through their corporate network. Often, however, customers do not allow outside service personnel to connect to their corporate network. As a result, you may need to access these interfaces through the HCP back-end network. This appendix provides instructions for doing this.
Accessing HCP interfaces through the back-end network Installing and Maintaining an HCP System
E1
However, for you to access the Tenant Management Console, HCP needs to know the fully qualified hostname of the applicable tenant. This is true regardless of whether you access the Tenant Management Console directly by entering its URL in your web browser address field or indirectly by clicking on its URL in the System Management Console. Similarly, for you to access the Namespace Browser, HCP needs to know the fully qualified hostname of the applicable namespace. This is true regardless of whether you access the Browser directly by entering its URL in your web browser address field or indirectly by clicking on its URL in the Tenant Management Console. Because the back-end network does not support DNS, the browser cannot resolve these URLs to IP addresses. Therefore, for each applicable tenant or namespace, you need to provide a mapping of the tenant or namespace hostname to the back-end IP address of an available storage node in the HCP system. You specify these mappings in the hosts file on your laptop computer:
c:\windows\system32\drivers\etc\hosts
E2
Accessing HCP interfaces through the back-end network Installing and Maintaining an HCP System
Each line in the hosts file must be a mapping of a fully qualified hostname to a back-end IP address. For example, if the back-end IP address of one of the nodes in the HCP system named hcp-ma.example.com is 10.1.1.16, you would add this line to the hosts file to enable access to the Tenant Management Console for the Finance tenant in that system:
10.1.1.16 finance.hcp-ma.example.com
Similarly, if the Finance tenant has a namespace named AccountsReceivable, you would add this line to the hosts file to enable access to the Namespace Browser for that namespace:
10.1.1.16 accounts-receivable.finance.hcp-ma.example.com
You can use the same or different IP addresses in your hostname mappings. You can include comments in a hosts file either on separate lines or following a mapping on the same line. Each comment must start with a number sign (#). Blank lines are ignored. Tip: You may need to make frequent changes to the hosts file. For convenience, consider putting a shortcut to the file on your desktop.
Accessing HCP interfaces through the back-end network Installing and Maintaining an HCP System
E3
E4
Accessing HCP interfaces through the back-end network Installing and Maintaining an HCP System
Glossary
A
access protocol
See namespace access protocol.
alert
A graphic that indicates the status of some particular element of an HCP system in the System Management Console.
B
bond
A pair of ports that share an IP address.
C
capacity
The total amount of storage space in HCP, excluding the space required for system overhead and the operating system. This is the amount of space available for all data in the repository, including the fixed-content data, metadata, and any redundant data required to satisfy namespace DPLs.
Glossary1
Installing and Maintaining an HCP System
CIFS
Common Internet File System. One of the protocols HCP uses to provide access to the contents of the default namespace. CIFS lets Windows clients access files on a remote computer as if they were part of the local file system.
compliance mode
The retention mode in which objects under retention cannot be deleted through any mechanism. This is the more restrictive retention mode.
compression service
The HCP service that compresses object data, thereby freeing space for the storage of additional objects.
core software
See HCP core software.
cross-mapping
With external storage, the mapping of two sets of logical volumes to two nodes such that each set maps to both nodes.
custom metadata
One or more user-defined properties that provide descriptive information about an object. Custom metadata, which is normally specified as XML, enables future users and applications to understand and repurpose object content.
Glossary2
Installing and Maintaining an HCP System
domain
D
data access path
With external storage, the path established by the mapping of a logical volume to a node.
data migration
The process of moving data off devices that are being retired from an HCP system into free storage on the remaining devices. The devices involved can be either storage nodes in an HCP RAIN system or storage arrays in an HCP SAIN system.
data outage
A situation in which one or more data objects are inaccessible (for example, due to multiple concurrent node failures).
default namespace
A namespace that supports the HTTP, WebDAV, CIFS, NFS, SMTP, and NDMP protocols and does not require user authentication for data access. An HCP system can have at most one default namespace.
disposition service
The HCP service that automatically deletes expired objects.
distributor key
The name that determines the SSH keys required for access to the HCP nodes when theyre locked down.
DNS
See domain name system (DNS).
domain
A group of computers and devices on a network that are administered as a unit.
Glossary3
Installing and Maintaining an HCP System
DPL
See data protection level (DPL).
duplicate elimination
The process of transparently eliminating redundant data associated with two or more identical objects.
dynamic DPL
A namespace data protection level that, at any given time, matches the system-level DPL setting.
E
enterprise mode
The retention mode in which these operations are allowed:
Privileged delete Changing the retention class of an object to one with a shorter duration Reducing retention class duration Deleting retention classes
F
fixed-content data
A digital asset ingested into HCP and preserved in its original form as the core part of an object. Once stored, fixed-content data cannot be modified.
Glossary4
Installing and Maintaining an HCP System
G
garbage collection service
The HCP service that deletes data and metadata left in the repository by incomplete operations, thereby freeing space for the storage of additional objects.
H
hash value
See cryptographic hash value.
HCP
See Hitachi Content Platform (HCP).
HCP namespace
A namespace that requires user authentication for data access. An HCP system can have multiple HCP namespaces.
HCP node
See node.
Glossary5
Installing and Maintaining an HCP System
HCP service
HCP service
See service.
HCP Setup
The program you use to install, upgrade, and add nodes to an HCP system.
HCP software
The HCP core software together with the HCP search facility software. The HCP software does not include the Appliance Operating System.
HDDS
See Hitachi Data Discovery Suite (HDDS).
hotfix
An update to an HCP system that resolves a particular problem. Hotfixes are designed to resolve customer problems on a one-off basis.
HTTP
Hypertext Transfer Protocol. One of the protocols HCP uses to provide access to the contents of a namespace. Additionally, the HCP System Management, Tenant Management, and Search Consoles use HTTP to communicate with Console clients.
Glossary6
Installing and Maintaining an HCP System
metadata
HTTPS
HTTP with SSL security. See HTTP and SSL.
I
index
See search index.
indexing service
The HCP service that builds and maintains the search index stored on the search nodes in HCP.
L
link
See replication link.
lockdown
The condition of an HCP node wherein login requires the use of SSH with specific keys.
logical volume
A unit of storage. For internal storage, a logical volume corresponds to a disk partition. For external storage, it corresponds to one or more segments of one or more disks.
M
management API
See HCP management API.
metadata
System-generated and user-supplied information about an object. Metadata is stored as an integral part of the object it describes, thereby making the object self-describing.
Glossary7
Installing and Maintaining an HCP System
migration service
The HCP service that migrates data off selected storage nodes in an HCP RAIN system or selected storage arrays in an HCP SAIN system in preparation for retiring those devices.
multipathing
With external storage, multiple means of access to a logical volume from a single node.
N
namespace
A logical partition of the objects stored in an HCP system. A namespace consists of a grouping of objects such that the objects in one namespace are not visible in any other namespace. Namespaces are configured independently of each other and, therefore, can have different properties.
NDMP
Network Data Management Protocol. The protocol HCP supports for backing up and restoring objects in the default namespace.
NFS
Network File System. One of the protocols HCP uses to provide access to the contents of the default namespace. NFS lets clients access files on a remote computer as if they were part of the local file system.
node
A server running HCP software and networked with other such servers to form an HCP system.
Glossary8
Installing and Maintaining an HCP System
protection service
O
object
For a data object, an exact digital representation of data as it existed before it was ingested into HCP, together with the system and custom metadata that describes that data. A data object is handled as a single unit by all transactions and services, including shredding, indexing, versioning, and replication. For a directory or symbolic link in the default namespace, the digital representation of its metadata.
offline upgrade
An upgrade of an HCP system during which the system is automatically shut down and then restarted when the upgrade is complete. During an offline upgrade, the system is not accessible to users and applications.
online upgrade
An upgrade of an HCP system during which the system remains online and accessible to users and applications.
P
permission
The granted ability to access the HCP System Management Console, Tenant Management, or Search Console and to perform a specific activity or set of activities in that Console.
ping
A utility that tests whether an IP address is accessible on the network by requesting a response from it. Also, to use the ping utility.
policy
One or more settings that influence how transactions and services work on objects. Such a setting can be a property of an object, such as retention, or a property of a namespace, such as versioning.
protection service
The HCP service that ensures the stability of the repository by maintaining a set level of data redundancy, as specified by the data protection level (DPL) for each namespace.
Glossary9
Installing and Maintaining an HCP System
protection set
protection set
A group of DPL storage nodes in which HCP tries to store all the copies of a given object.
protocol
See namespace access protocol.
Q
query API
See metadata query API.
R
RAIN
Redundant array of independent nodes. An HCP system configuration in which the nodes use internal or direct-attached storage.
region
An HCP-internal logical grouping of objects and certain types of system data (such as system log messages and usage statistics). HCP uses regions to distribute object metadata and system data among the storage nodes in the system.
replica
The target system to which the replication service copies objects and other information from the primary system during replication.
replication
The process of keeping selected tenants and namespaces in two HCP systems in sync with each other. Basically, this entails copying object creations, deletions, and metadata changes from one system to the other. HCP also replicates tenant and namespace configuration, data access accounts, retention classes, all compliance log messages, and all HCP tenant log messages.
replication link
A configurable association that specifies the HCP systems from and to which the replication service should copy objects and other data.
replication service
The HCP service that performs replication.
Glossary10
Installing and Maintaining an HCP System
Search Console
repository
The aggregate of the namespaces defined for an HCP system.
retention mode
A namespace property that affects which operations are allowed on objects under retention. A namespace can be in either of two retention modes: compliance or enterprise.
retention period
The period of time during which an object cannot be deleted.
retention setting
The property that determines the retention period for an object.
role
A named collection of permissions that can be associated with an HCP user account, where each permission allows the user to perform some specific interaction or set of interactions with the HCP System Management Console, the Tenant Management Console, or, for default namespaces only, the Search Console. Roles generally correspond to job functions.
S
SAIN
SAN-attached array of independent nodes. An HCP system configuration in which the nodes use SAN-attached storage.
scavenging service
The HCP service that ensures that objects in the HCP repository have valid metadata.
Search Console
The web application that provides interactive access to the search functionality of the active search system.
Glossary11
Installing and Maintaining an HCP System
search facility
search facility
An interface between the search functionality provided by a system such as HDDS or HCP and the HCP Search Console. Only one search facility can be enabled at any given time.
search index
An index of the metadata and key terms in namespace objects. The active search system builds, maintains, and stores this index.
search node
An HCP node that runs the HCP search facility software and stores the search index thats built and maintained by HCP.
service
A background process that performs a specific function that contributes to the continuous tuning of the HCP system. In particular, services are responsible for optimizing the use of system resources and maintaining the integrity and availability of the data stored in the HCP repository.
shredding
The process of deleting a data object and overwriting the locations where its bytes were stored in such a way that none of its data or metadata can be reconstructed. Also called secure deletion.
shredding service
The HCP service that shreds deleted objects that are marked for shredding.
SMTP
Simple Mail Transfer Protocol. The protocol HCP uses to receive and store email data directly from email servers.
SNMP
Simple Network Management Protocol. A protocol HCP uses to facilitate monitoring and management of the system through an external interface.
Glossary12
Installing and Maintaining an HCP System
SSH
Secure Shell. A network protocol that lets you log into and execute commands in a remote computer. SSH uses encrypted keys for computer and user authentication.
SSL
Secure Sockets Layer. A key-based Internet protocol for transmitting documents through an encrypted link.
storage node
An HCP node that runs the complete HCP software (except the HCP search facility software) and stores the objects added to HCP.
subdomain
A subset of the computers and devices in a domain.
syslog
A protocol used for forwarding log messages in an IP network. HCP uses syslog to facilitate system monitoring through an external interface.
system metadata
System-managed properties that describe the content of an object. System metadata includes policies, such as retention and data protection level, that influence how transactions and services affect the object.
Glossary13
Installing and Maintaining an HCP System
tenant
T
tenant
An administrative entity created for the purpose of owning and managing namespaces and data access accounts. Tenants typically correspond to customers, business units, or individuals.
U
Unix
Any UNIX-like operating system (such as UNIX itself or Linux).
user account
A set of credentials that gives a user access to one or more of the System Management Console, Tenant Management Console, HCP management API, metadata query API, and Search Console.
user role
See role.
V
volume
See logical volume.
W
WebDAV
Web-based Distributed Authoring and Versioning. One of the protocols HCP uses to provide access to the contents of the default namespace. WebDAV is an extension of HTTP.
WORM
Write once, read many. A data storage property that protects the stored data from being modified or overwritten.
Glossary14
Installing and Maintaining an HCP System
zero-copy failover
Z
zero-copy failover
The process of one node automatically taking over management of storage previously managed by another node that has become unavailable.
Glossary15
Installing and Maintaining an HCP System
zero-copy failover
Glossary16
Installing and Maintaining an HCP System
Index
A
access protocols, namespace 1-3 accessing HCP interfaces through back-end network E-1E-3 Active Directory server time synchronization 2-6 active search system 1-3 adding chassis IP addresses 9-7 nodes 5-25-13 SSH keys in Linux D-3D-4 SSH keys in Windows D-2D-3 storage 7-27-10 administrative interface for HCP search facility, accessing 10-6 advanced alerts 11-2 advanced options for installing HCP software 3-263-27 alerts advanced 11-2 upgrade completion 4-26 AMS-200 2-7 Appliance Operating System 1-10 See also operating system Appliance OS Installation Information form A-2 arc_nodes configuration file section C-2 backup during online upgrade 4-4 BIOS date 3-3 blade servers See CB 320 servers; nodes blade_servers configuration file parameter C-3
C
C components log files 11-611-7 logging levels, changing 11-911-10 canceling region-count changes 9-139-16 capablemsgversion configuration file parameter C-3 CB 320 servers See also nodes about 1-8 adding nodes 5-4 adding/removing chassis IP addresses 9-7 configuring for HCP 2-7 CD trays, ejecting 11-3311-34 chained writes, disabling 10-510-6 Change Password page 3-22 changing DNS settings 9-29-4 HTTP persistent connection timeout interval 9-10 install user password 3-133-14 logging levels 11-811-12 network settings 9-49-7 region count 9-139-16 security account password 3-22 serial numbers 9-99-10 time settings 9-89-9 chassis See CB 320 servers chassis_ip_csl configuration file parameter C-4 checking space for upgrading 4-74-13
B
back end access to HCP interfaces E-1E-3 IP addresses for new nodes 5-7 MTU, changing 9-49-7 network 2-3 node IP addresses, about 2-2 node IP addresses, changing 9-49-7 node IP addresses, setting 3-6 back-end network 1-6 backend_mtu configuration file parameter C-3
Index1
Installing and Maintaining an HCP System
D
data access accounts about 1-3 creating/configuring during online upgrades 4-3 data migrations about 8-2 finalizing 8-28-9 upgrades during 4-2, 8-2 data outages during data migration 7-12 during online upgrade 4-4 database, recovering 6-4, 6-11 default namespace 1-2 default system configuration, resetting 3-9 default tenant 1-2 deleting encryption keys for log downloads 11-14 deletion records, enabling retention of 2-10 disabling chained writes 10-510-6 DPL one 9-119-12 encryption 3-19 protection service for storage addition 7-4 replication after installation 9-17 replication during installation 2-8, 3-18 services 10-210-3 zero-copy failover after installation 9-10 9-11 zero-copy failover for TrueCopy data migration 7-13 disk_flush configuration file parameter C-4 distributor key about 2-3 setting 3-17 distributor_key configuration file parameter C-4 DNS about 2-42-5 adding new nodes as master servers 5-3 dummy name for HCP system 2-5 settings, changing 9-29-4 settings, setting 3-17 DNS Settings page 9-39-4 dns_enabled configuration file parameter C-5
Index2
Installing and Maintaining an HCP System
getProtectionSets.action page
dns_servers configuration file parameter C-5 domain name system See DNS downloading HCP internal logs encryption keys for 11-1311-14 procedure 11-1211-13 DPL configuring system-level setting 2-10 enabling/disabling DPL one 9-119-12 dummy DNS name for HCP system 2-5 HCP software installation 3-153-16, 4-19 4-22, 5-95-11, 6-76-9, 6-146-16, 6-226-23, 7-57-7 node list 3-273-31 filesystem configuration file parameter C-6 finalizing data migrations procedure 8-48-9 RAIN systems 8-28-3 requirements 8-38-4 SAIN systems 8-3 finding user IDs 11-5 fixed-content storage system 1-2 forms Appliance OS Installation Information A-2 Encryption Key A-5 HCP Node Addition Information A-6 HCP Software Installation Information A-3 A-4 forwarding SSH keys Linux D-3D-4 Windows D-2D-3 freeing space on nodes 4-144-16 front end IP addresses for new nodes 5-7 MTU, changing 9-49-7 network IP address, about 2-2, 2-3, 2-4 network IP address, changing 9-49-7 network IP address, setting 3-5, 3-17 node IP addresses, about 2-2 node IP addresses, changing 9-49-7 node IP addresses, setting 3-5 subnet mask 2-2 front-end network 1-6 frontend_mtu configuration file parameter C-6
E
ejecting CD trays 11-3311-34 enable_replication configuration file parameter C-5 enabling DPL one 9-119-12 encryption 3-19 Management API 2-10 replication after installation 9-169-17 replication during installation 2-8, 3-18 retention of deletion records 2-10 services 10-210-3 zero-copy failover after installation 9-10 9-11 encryption about 2-82-9 alert 11-2 configuration file parameter C-5 enabling/disabling 3-19 key, saving during installation 2-82-9, 3-20 keys for downloading logs 11-1311-14 Encryption Key form A-5 encryption_cipher configuration file parameter C-5 encryption_ivmode configuration file parameter C-6 encryption_key_length configuration file parameter C-6 external storage AMS-200 2-7 architecture 1-71-8 See SAN-attached storage WMS-100 2-7 external time source 2-52-6
G
gateway configuration file parameter C-6 gateway router IP address, about 2-2, 2-3, 2-4 IP address, changing 9-49-7 IP address, setting 3-17 gathering information about 2-1 HCP software installation 2-32-9 HCP upgrade 4-6 node addition 5-5 node recovery 6-36-4 operating system installation 2-22-3 post-installation setup 2-92-10 storage addition 7-3 getAdminView.action page 11-32 getMM.action page 11-33 getProtectionSets.action page 11-33
F
failed upgrades, restarting 4-274-28 fibre channel switches 1-8 files cluster.cfg 3-323-33, C-1C-9
Index3
Installing and Maintaining an HCP System
getUUIDs.action page
getUUIDs.action page 11-33 encryption key during installation 2-82-9, 3-20 information for installing 2-32-9 installation files, retrieving 3-153-16, 4-194-22, 5-95-11, 6-76-9, 6-14 6-16, 6-226-23, 7-57-7 installation requirements 3-113-12 installing 3-123-20 installing, advanced options 3-263-27 verifying installation 3-203-25 HCP Software Installation Information form A-3 A-4 HCP system about 1-21-4 alerts 11-2 configuration, viewing C-1 configuring 3-73-9, 3-163-19 customer support contact information 2-8 DNS settings 2-42-5 encryption 2-82-9 hardware 1-41-8 hotfixes, installing 4-324-36 initial time 2-6 installation, verifying 3-203-25 installing 3-23-20 managing 10-1 monitoring during online upgrades 4-5 nodes 2-9 post-installation configuration 3-26 preparing for storage addition 7-4 preparing to upgrade 4-16 reconfiguring 9-19-2 replication 2-8 serial number 2-7 software 1-101-11 storage array migration 7-117-18 storage configuration 2-7 time source 2-52-6 time zone 2-6 troubleshooting 11-1 upgrade, verifying 4-244-25 upgrading 4-174-27 upgrading, about 4-24-3 HCP System Startup page 3-22 HDDS search facility 1-3 See also search facilities highest-numbered node 2-3 Hitachi Content Platform See HCP system Hitachi Device Manager integration, enabling 2-9 hostname prefix 9-29-4 hostname_prefix configuration file parameter C-6
H
hardware dependent HCP features 1-91-10 networking 1-6 nodes 1-5 overview 1-4 HCP back-end network 1-6 front-end network 1-6 hardware 1-61-8 HCP 300 1-4 HCP 500 1-4 HCP core software 2-1 See also HCP software HCP internal logs daily logs 11-7 downloading 11-1211-13 encryption keys for downloading 11-13 11-14 log files 11-611-7 log rotation 11-711-8 logging levels, about 11-811-9 logging levels, changing for C components 11-911-10 logging levels, changing for JVM 11-10 11-12 managing 11-511-6 HCP namespaces 1-2 HCP Node Addition Information form A-6 HCP search facility See also search facilities about 1-3 administrative interface, accessing 10-6 HCP Setup See also HCP Setup wizard about 3-7 adding storage 7-2, 7-87-10 logs 11-3 HCP Setup wizard See also HCP Setup adding nodes 5-115-13 configuration options 3-83-9, 3-173-19 configuring HCP system 3-163-19 installing HCP 3-163-20 menus B-1B-5 navigating 3-73-8 resetting configuration options 3-9 saving configuration settings 3-9 upgrading HCP 4-224-27 HCP software core 2-1
Index4
Installing and Maintaining an HCP System
J
JVM log files 11-7 logging levels, changing 11-1011-12 snapshot, creating 11-1411-15
I
indexing during online upgrade 4-4 install user password, changing 3-133-14 password, resetting 11-311-4 installation files, HCP software retrieving for HCP installation 3-153-16 retrieving for HCP upgrade 4-194-22 retrieving for node addition 5-95-11 retrieving for node recovery 6-76-9, 6-14 6-16, 6-226-23 retrieving for storage addition 7-57-7 installation SSH key D-1 installing HCP software 3-123-20 HCP software, advanced options 3-263-27 hotfixes 4-324-36 operating system 3-43-6 operating system for logical-volume recovery 6-12 operating system for OS recovery 6-6 operating system for search node recovery 6-21 operating system on new nodes 5-8 using modified configuration file 3-323-35 using node-list files 3-273-31 internal logs See HCP internal logs internal storage architecture 1-61-7 internal time source 2-5 IP addresses back end, about 2-2 back end, changing 9-49-7 back end, setting 3-6 front end, changing 9-49-7 front end, setting 2-2, 3-5 front-end network, about 2-4 front-end network, changing 9-49-7 front-end network, setting 2-2, 2-3, 3-5, 3-17 multicast, about 2-4 multicast, changing 9-49-7 multicast, setting 3-17 IP addresses for new nodes 5-7
K
kerberos_ssh configuration file parameter C-6 keys encryption 2-82-9 SSH D-1
L
links, resetting checkpoints 10-410-5 Linux, adding and forwarding SSH keys D-3D-4 listing encryption keys for log downloads 11-14 lockdown configuration file parameter C-7 logging in administrator interface for HCP search facility 10-6 HCP software installation 3-123-14 HCP upgrade 4-174-19 migration finalization 8-5 node addition 5-8 node recovery 6-6, 6-13, 6-21 storage addition 7-47-5 System Management Console 3-21 logical volumes adding 7-27-10 cross-mapping 1-91-10 multipathing 1-9 recovering 6-126-18 viewing on getUUIDs.action page 11-33 login page 3-21 logs See also HCP internal logs HCP Setup 11-3 server SEL 11-6, 11-8
M
Management API, enabling 2-10 managing HCP internal logs 11-511-6 HCP system 10-1 services 10-2 max_simul_disks configuration file parameter C-7 max_simul_nodes configuration file parameter C-7 mcast configuration file parameter C-7 mdpl configuration file parameter C-7
Index5
Installing and Maintaining an HCP System
N
Namespace Browser, access through back-end network E-2E-3 namespaces about 1-2 access protocols 1-3 creating/configuring during online upgrades 4-3 navigating HCP Setup wizard 3-73-8 NDMP 1-3 Network Settings page 9-59-6, 9-7 network settings, changing 9-49-7 networking about 1-6 installation information 2-4 replication 1-81-9 NFS chained writes, disabling 10-510-6 mounts during online upgrade 4-34-4 support 1-3 node addition about 5-25-3 information for 5-5 IP addresses for new nodes 5-7 procedure 5-75-13 verifying 5-135-14 node recovery search, procedure 6-216-25 search, verifying 6-25 storage, information for 6-36-4
O
objects about 1-2 storage 1-5 offline upgrades See also upgrading about 4-3 rolling back 4-294-32
Index6
Installing and Maintaining an HCP System
P
Pageant D-2 passwords administrative interface for HCP search facility 10-6 install user, changing 3-133-14 install user, resetting 11-311-4 security account, changing 3-22 persistent connection timeout interval, changing 9-10 ping, disabling 2-9 policy_enabled configuration file parameter C-8 post-installation configuration 3-26 setup information 2-92-10 prechecks for upgrading 4-74-13 preparing HCP system for storage addition 7-4 HCP system for upgrade 4-16 protection service disabling for storage addition 7-4 during online upgrade 4-4 reenabling after storage addition 7-10 protection sets, viewing current 11-33 protocols, namespace access 1-3 PuTTY 3-11, D-2D-3
R
RAIN systems about 1-4 architecture 1-61-7
Index7
Installing and Maintaining an HCP System
S
SAIN systems See also SAN-attached storage about 1-4 architecture 1-71-8 fibre channel switches 1-8 SAN-attached storage See also SAIN systems array migration 7-117-18 configuration review 3-33-4 retiring arrays 8-2, 8-3 shared 2-7 virtualizing 7-157-17 saving system configuration settings 3-9 scp 11-13 screen command 3-10 Search Console See also search facilities about 1-3 use during online upgrades 4-4 search facilities See also search index; search nodes about 1-31-4 enabling 2-10 search index See also search facilities; search nodes about 1-3
Index8
Installing and Maintaining an HCP System
upgrading
adding and forwarding in Windows D-2D-3 opening install shell 3-10 SSH, enabling/disabling 2-10 SSL server certificate, new 2-10 starter account 1-11 starting services 10-210-3 stopping services 10-210-3 storage adding 7-27-10 reviewing SAN-attached configuration 3-3 3-4 shared SAN-attached 2-7 type, setting 3-18 types 1-5, 2-7 verifying addition 7-107-11 storage addition information for 7-3 preparing HCP system 7-4 procedure 7-37-10 verifying 7-107-11 storage arrays See SAN-attached storage Storage Node page alerts 11-2 removing nodes 8-11 shutting down nodes 8-11 storage nodes about 1-5 recovering 6-2 retiring in RAIN systems 8-1 retiring in SAIN systems 8-1 verifying recovery 6-18, 6-25 storage_type configuration file parameter C-9 subnet back end, about 2-2 back end, changing 9-5 mask, front end 2-2 masks, changing 9-49-7 masks, setting 3-5 suspending replication for HCP upgrade 4-16 syslog logging, enabling 2-10 system health, verifying after installation 3-23 3-25 System Management Console about 1-11 access through back-end network E-2 alerts 11-2 logging in 3-21 restricting access 2-10 verifying HCP installation 3-213-23 verifying HCP upgrade 4-244-25 verifying node addition 5-135-14 verifying node recovery 6-18, 6-25 verifying storage addition 7-107-11 system service tools about 11-1511-16 cluster_get 11-2411-29 cluster_put 11-2911-32 cluster_ssh 11-1711-24 output 11-17 running 11-16 target nodes 11-1611-17 target system 11-16 system status pages, viewing 11-3211-33 system-level DPL setting 2-10, 9-119-12
T
Tenant Management Console about 1-11 access through back-end network E-2E-3 tenants about 1-2 creating/configuring during online upgrades 4-3 Tiered Storage Manager about 7-11 migrating data 7-18 time settings, about 2-52-6 settings, changing 9-89-9 settings, setting 3-18 Time Settings page 9-89-9 timeserver configuration file parameter C-9 timezone configuration file parameter C-9 troubleshooting 11-1 TrueCopy about 7-11 migrating data 7-127-15 type configuration file parameter C-2
U
updating nodes to recognize new storage arrays 7-147-15 Upgrade page 4-26 upgrading about 4-2 committing upgrades 4-254-27 completion alert 11-2 during data migration 4-2, 8-2 HCP system 4-174-27 information for 4-6 in-progress alert 11-2 offline 4-3
Index9
Installing and Maintaining an HCP System
upgrading, online
online 4-34-5 prechecks 4-74-13 preparing HCP system 4-16 requirements 4-54-7 restarting failed upgrades 4-274-28 rolling back offline upgrades 4-294-32 space for 4-74-16 uploading encryption keys for log downloads 11-14 URLs administrator interface for HCP search facility 10-6 System Management Console 3-21 user accounts about 1-11 creating 2-10 creating/configuring during online upgrades 4-3 security 1-11 user roles 1-11 using HCP Setup wizard 3-73-9 zero-copy failover about 1-91-10 alerts 11-2 disabling for logical volume recovery 6-12 disabling for storage reconfiguration 7-13 enabling/disabling after installation 9-10 9-11 during online upgrade 4-4 reenabling after logical-volume recovery 6-18 reenabling after storage reconfiguration 7-15
V
verifying HCP installation 3-203-25 HCP upgrade 4-244-25 hotfix installation 4-36 node addition 5-135-14 offline upgrade rollback 4-32 serial number 3-21 storage addition 7-107-11 storage node recovery 6-18, 6-25 viewing current node status 11-32 current protection sets 11-33 current regions 11-33 HCP system configuration C-1 logical volume information 11-33 system status pages 11-3211-33 virtual terminal session 3-93-11 virtualizing SAN-attached storage 7-157-17
W
WebDAV 1-3 Windows, adding and forwarding SSH keys D-2 D-3 WMS-100 2-7 WORM 1-2
Z
zero_copy_failover configuration file parameter C-9
Index10
Installing and Maintaining an HCP System
Hitachi Data Systems Corporate Headquarters 750 Central Expressway Santa Clara, California 95050-2627 U.S.A. Phone: 1 408 970 1000 www.hds.com info@hds.com Asia Pacific and Americas 750 Central Expressway Santa Clara, California 95050-2627 U.S.A. Phone: 1 408 970 1000 info@hds.com Europe Headquarters Sefton Park Stoke Poges Buckinghamshire SL2 4HD United Kingdom Phone: + 44 (0)1753 618000 info.eu@hds.com
FE-98ARC018-07