Sie sind auf Seite 1von 8

Date: January 24, 2012

VNX Issues and Recommended Actions


The following table (which can also be found in EMC Knowledgebase article emc271233) is a list of
known VNX issues or topics and workarounds as well as recommended solutions, as applicable.
Important: This table lists the latest Issue first, then the next latest, and so on. This list was accurate at
publication time but may have since been updated. Always check EMC Knowledgebase article
emc271233 and its referenced Knowledgebase articles for the latest information.
Note: The EMC Knowledgebase articles in the Solution column of this table provide detailed information
about the Issue and its solution. To read an article, go to the EMC Online support web page, click the
Search Support tab, and then enter the Knowledgebase article number. Next, select Solutions from the
drop-down menu, and then click Search.
Issue

Description

Workaround

Solution

Fix

Conducting
EFD drive
firmware
updates on all
new VNX
installations
per FCO
F011912FOEFD.

All new VNX Unified/File and Block


installations require that EFD drive
firmware be updated on affected
drive types, prior to configuring the
storage system.

See emc286025 and refer to


F011912FO-EFD for more details.

emc286025

TBD

Unisphere
inconsistently
displays
(missing) File
Events from
the File Event
Logs.

An enhancement to the size and


number of File Event Logs introduced
a problem with consistently displaying all File Events in Unisphere.

See emc284381 for a workaround


fix for this issue.

emc284381

TBD

Managing Dual
Solid State
Drive (SSD)
issues for VNX
and CLARiiON
CX4 systems.

It has been found that once single


SSD soft media errors begin
registering within a RAID Group/Pool,
the array may not be able to disable
the faulting drive before a 2nd drive
begins experiencing the same errors,
resulting in a potential DU situation.

Review ETA emc282741 for a


summary of the issues and
workarounds.

ETA emc282741

TBD

See emc266585 to disable FAST


Cache.
See emc283576 to find 0x820
errors in SP logs.
See emc284440 to setup

-1-

VNX Issues and Recommended Actions


Issue

Description

Workaround

Solution

Fix

emc273267

N/A

monitoring of SSD drives.


See emc2844355 for how the
Support Center should handle
cases where EFDs dial home with
error code 820 Soft Media Errors.

Cannot change
SP IP address
from
Unisphere

Cannot change SP IP address from


Unisphere management console on
VNX Unified/File systems.

Beginning with
7.0.35.3/05.31.000.502
and later, you can no longer
change the SP IP address from the
GUI. This is by design.
This issue does not apply to VNX
Block-only systems.

File alerts no
longer appear
in Unisphere.

Due to an event parsing issue, File


events are not posted to Unisphere
after October 30, 2011. This issue
only affects Alerts/Events posted to
Unisphere, and does not affect the
ability of the system to CallHome for
applicable issues.

See emc281631 for workaround fix


for this issue.

ETA emc281631

Fixed in VNX OE
for File 7.0.50.2

VNX IP
Address
change procedure on VNX
OE
7.0.35.3/05.31
.000.5.502 and
later

Changing VNX IP addresses on VNX


OE 7.0.35.3/05.31.000.5.502 and
later.

See emc279637 for steps to


change IP addresses on SPs and
Control Stations.

emc279637

N/A

Replication
checkpoints
can become
corrupt and
lead to Data
Mover panics

There is a regression issue with


7.0.35.3, where a file systems
replication checkpoints can become
corrupt if F-RDE deduplication is
enabled, causing the Blade to panic.

Please contact your Service


Provider for assistance regarding
this issue.

ETA emc277607

VNX for File OE


patch 7.0.35.301
will prevent
server panic.

Adding or
Removing VNX
systems to
local domains

This knowledgebase shows how to


add or remove a VNX system from a
Local domain with VNX OE
7.0.35.3/05.31.000.5.502.

See emc269625 to change VNX IP


addresses prior to 7.0.35.3.

Fixed in VNX for


File OE 7.0.40.1.
See emc278385 for Unisphere
examples of adding or removing a
system from the Local domain, as
well as what happens to domain
accounts in the system being
added or removed.

-2-

emc278385

N/A

VNX Issues and Recommended Actions


Issue

Description

Workaround

Solution

Fix

VNX
Unified/File
systems do
not call home
for all Block
events.

The Unified/File Control Station


template file is outdated and does
not properly flag and dispatch all
VNX Block CallHome events for
CallHome.

See ETA emc277773 for more


information and a workaround for
the issue.

ETA emc277773

Fixed in
VNX for File OE
7.0.40.1

NDU of Block
Enablers can
disable READ
cache and
reduce Write
cache values

Certain array enablers require


dedicated memory resources. An
NDU of certain enablers will forcibly
reduce Read and Write cache values,
and disable READ cache.

See emc277689 for an example of


an Enabler NDU that changes
cache values and disables READ
cache.

emc277689

N/A

SP POST
health check
upgrade
failure
message

Both File OE and Block OE upgrade


checks will flag SP POST versions
36.80 or below, and block the
upgrade process.

See emc277473 for more


information related to the FCO
F041211FC, and bad POST
versions on the SPs.

emc277473

FCO F041211FC

Domain may
become unfused and
broken

The VNX domain may break after


changing the Control Stations IP
address, or conducting Control
Station failover operations.

See emc277605 for more


information on this issue, which
points to the appropriate
knowledgebase solution to re-fuse
the domain.

emc277605

N/A

Configuring an
IPalias for dual
VNX Control
Stations

With the release of VNX OE


7.0.35.3/05.31.000.5.502, the
Control Station IPalias feature has
been restored.

See emc277437 for an example of


how to configure an IPalias using
dual VNX Control Stations,
including Control Station failover
and failback.

emc277437

Corrected in VNX
for File OE
7.0.35.3.

Disabling the
Unisphere
workaround
for root access
to Control
Station
functions

With the release of VNX OE


7.0.35.3/05.31.000.5.502, the root
user workaround for Unisphere
management of the Control Station is
no longer required.

See emc270273 to disable the


workaround that was published
with emc258105.

emc270273

Corrected in VNX
for File OE
7.0.35.3.

SAS Driver
issue leads to
Drive Errors &
SP Panics

Multiple SP Bug check issues with the


SAS driver interaction and the disk
drives can lead to drive errors,
timeouts, and eventual SP panics.

No workaround.

emc271461

Fixed in Block
OE patch
05.31.000.5.011.

-3-

VNX Issues and Recommended Actions


Issue

Description

Workaround

Solution

Fix

CRITICAL
alerts and
Callhomes for
mismatched
OE versions

When upgrading from 7.0.12.07.0.14.0 and 05.31.000.5.006-012 OE


versions, to the OE versions
[7.0.35.3/05.31.000.5.502], the
system will Callhome if both the File
and Block are not updated to OE
7.0.35.3/05.31.000.5.502 within 4
hours.

See emc276260 for more details.


Essentially the system will
Callhome after the File/Block OE
versions become mismatched after
4 hours, then every 48 hours
thereafter until corrected.

emc276260

N/A

Setup of SP IP
addresses
may fail
during VIA
initialization.

During the VIA initialization process,


occasionally the SP IP setup fails
with: Backend IP and Name
Setup | Could not change IP
address of the SP.

Follow the guidance found in


emc275841 to resolve the failure
and complete the VIA setup.

emc275841

Corrected in VNX
for File OE
7.0.35.3.

Configuring
domain
system
account on
VNX.

Upgrades to VNX
7.0.35.3/05.31.000.5.502 requires
manual configuration of the System
Account attribute by default.

See emc270977 for the procedure


to configure the System Account
with VNX OE for File 7.0.35.37.0.40.1 (Block OE
05.31.000.5.502-.509).

emc270977

Fixed on VNX OE
7.0.50.2/05/31.0
00.5.704.

emc274779

N/A

Recommended
Cache Page
and
Watermark
settings

Note: Factory installs automatically


create the sysadmin user with the
System Account attribute by
default.

With 7.0.50.2/05.31.000.5.704, the


system account is automatically
assigned during the upgrade.

Default Cache Page and Watermark


settings may not be set out of the
factory in accordance with best
practice values.

See emc274779 for the


recommended Cache Page and
Watermark settings on the VNX
arrays.
See emc267304 for Best Practice
settings for Read and Write Cache.

Blade panics
from Mac OS X
Lion (10.7)
client access

Mac OS X Lion (10.7) CIFS clients


can panic the VNX Blades (Data
Movers) due to unexpected null
values, resulting
in Rolling Panics.

Stop access from all Mac OS X Lion


(10.7) clients, apply patch to stop
rolling panics.

emc263721

Apply patch to
stabilize, upgrade
to 7.0.13.1 or
later.

Control Station
558 Callhome
events when
Blades cabled
to SPS units

Blade enclosure power cables are


incorrectly plugged into SPS power
outlets on VNX5300/5500 Unified
systems, and weekly SP battery test
results in CRITICAL 558 Callhome
events.

Fix is to re-cable the Blade


Enclosure power cables correctly,
and remove from SPS power outlet
and plug into the rack power
distribution units (PDUs).

emc273879

Installation misconfiguration
issue.

Restarting the
VIA program.

Outlines the steps required to


re-enter the VIA initialization process
if the process was interrupted prior
to completing all initialization steps.

Refer to the solution for the


necessary steps to restart the VIA
when initializing the VNX
Unified/File system.

emc270909

N/A

-4-

VNX Issues and Recommended Actions


Issue

Description

Workaround

Solution

Fix

VNX for File


OE version vs.
VNX for Block
OE version
compatibility
table.

Table that displays compatible File


OE to Block OE versions.

Refer to this knowledgebase


solution when determining which
VNX OE versions are compatible
between File and Block
components.

emc261310

N/A

Unisphere 1.1
cannot be
used to
manage legacy
Celerra or
CLARiiON

The Unisphere 1.1 release cannot be


used to manage any Celerra or
CLARiiON legacy systems.

Use a Unisphere Client version


appropriate for the legacy system,
or use a web browser connection
to manage the system.

emc261017

Corrected in VNX
OE 7.0.35.3.
/05.31.000.5.502,
and Unisphere
1.1.25.1.01.29.

The Linux
Control
Station IP
alias feature
is disabled in
the first
release of the
VNX File OE.

The Control Station IP Alias feature


is not supported in the 7.0.12.0
release.

The IP alias feature is disabled on


the Linux Control Station. Cannot
use an IP alias for dual Control
Station environments at this time.

emc259516

VNX for File OE


7.0.35.3 reenables IPalias
feature for
Control Station.

Password
change to
VNX
administrator
account
causes
communication failure
between File
and Block
components.

After a Global administrator


password change, the File Control
Station can no longer communicate
with the Block backend.

Run nas_storage -m id=1 -s


to synchronize the administrator
password credentials to the
Control Station.

emc261195

Corrected in VNX
for File OE
7.0.35.3.

Unisphere
connection
error or login
error

Unisphere connection or login error


after using nas_config -ssl to
generate an SSL certificate

Regenerate the SSL certificate


using
/nas/sbin/nas_ca_certificate generate.

emc266560

Corrected in
VNX for File OE
7.0.13.1,
7.0.14.0, and
7.0.35.3.

Root user
workaround
for Unisphere
access to the
Control
Station.

A workaround is required that


provides Control Station root privilege
to a Storage domain administrator
account in order to manage Control
Station settings, ConnectHome, etc.

Set up a domain account and


assign root UID=0 privileges on
the Control Station for the
selected Storage administrator
user account.

emc258105

Corrected in VNX
for File OE
7.0.35.3. Root
User workaround
no longer
required. Login
as Root Scope
Local to manage
Control Station
properties.

-5-

VNX Issues and Recommended Actions


Issue

Description

Workaround

Solution

Fix

Recommended
cache values
for VNX arrays
on initial
system
installation.

Recommended cache settings on


VNX Block and Unified arrays.

Block systems ship without cache


enabled, and Unified systems set
cache defaults contrary to best
practices. Use the guidance
outlined in the knowledgebase
solution to set the basic Read and
Write cache values for the VNX
array.

emc267304

Installation
documentation
for Block-only
systems, and in
File OE 7.0.35.3
for Unified/File
systems.

Non-disruptive
Blade enclosure
fault LED and
alert.

Under certain conditions a fault LED


will be illuminated on the Blade
Enclosure and a CB x Csum Error
will appear in a System Alert.

Although the issue has no impact


to the system, to correct the
resume prom checksum issue,
contact your Service Provider.
The issue can be resolved without
disrupting user access.

emc267489

Corrected in VNX
for File OE
7.0.35.3.

VIA IPMI cable


check failure

VIA initialization with dual Control


Stations may fail with Secondary
Control Station is unreachable on
IPMI network.

The solution explains possible


workaround steps to get past the
issue.

emc269717

Corrected in VNX
for File OE
7.0.35.3.

VNX IP Address
Change
procedure
prior to
7.0.35.3/05.31.
000.5.502

Changing VNX IP addresses VNX OE


7.0.35.3/05.31.000.5.502 and later.

Contact your Service Provider for


assistance in changing Control
Station and/or SP IP addresses on
a production system.

emc269625

N/A

Dual Control
Station (CS)
operation with
VNX domains
and Unisphere.

Unisphere login works with either


CS0 or CS1 running as the Primary
Control Station, however, there is a
domain IP update process that
takes approximately 15 minutes to
complete after every Failover or
Failback operation, during which
Unisphere will be unable to log into
the VNX system.

Expected operation of dual Control


Stations with Unisphere.

emc260055

No fix, just an
explanation of
how Control
Station failover
and failback
works within the
Unisphere
domain.

Invalid Time
issue on Client

Unisphere login from Client system


fails with the "Certificate has invalid
date"

Set the Windows workstation PC


Clock ahead to a time slightly
before the "Valid From" date in
the details of the Certificate
(generally about 6 hours ahead)

emc265473

Environmental

Time issue
between SP
and Client

Time difference between client and


SPs creates certificate date problem,
preventing access to Unisphere

Follow the workaround outlined in


the Knowledgebase solution.

emc247504

TBD

See emc279637 for changing VNX IP


addresses on 7.0.35.3 OE and later.

Note: The Control Station and SP


IP addresses must always be on
the same subnet.

-6-

VNX Issues and Recommended Actions


Issue

Description

Workaround

Solution

Fix

Initialization
utility fails to
discover VNX
Block system

Unisphere Initialization Utility for


Block fails to discover Block systems
during installation.

If running Cisco Security: the


Cisco Security Agent may be
blocking the wizard. Set CSA
service to disabled, reboot
system, and retry Initialization
Utility.

emc265863

N/A

If running another Firewall


product, check if it is blocking
the Initialization Utility
application, disable and retry.
Invalid Time
Zone issue in
VIA

Certain time zone settings in VIA do


not work during the initialization
process.

Click Back in VIA until you reach


the time zone settings screen,
and then change the setting to a
Region/City time zone that
matches your location.

emc261627

Corrected in the
VNX for File OE
7.0.35.3.

Single VNX
Domains, no
Legacy
management

The first release of the VNX product,


and Unisphere version 1.1, does not
support legacy domains or adding
other VNX systems to a VNX domain.

The VNX Unisphere multi-domain


capability, and legacy domain
management, will be provided in
the next scheduled maintenance
release.

emc266655

Corrected in the
VNX OE
7.0.35.3\05.31.00
0.5.502, and
Unisphere
1.1.25.1.0129.

MirrorView
Support

MirrorView (both MV/S and MV/A)


between VNX and CX4, CX3, AX4-5 is
supported, but naviseccli must be
used to establish the connections.
No GUI support.

Must use CLI commands to setup


MirrorView.

emc262013

Corrected in VNX
for Block OE
05.31.000.5.002.

USM Dual
Control Station
Issue

USM > Software Install > Install


VNX OE fails with Dual Control
Station configuration is not
supported message on systems with
Dual Control Stations.

Dual Control Station upgrades are


not supported using USM, except
for the Install VNX for Block
wizard, which allows the User to
upgrade Block OE and add
Enablers/Firmware to the array.

emc266770

Dual CS support
with USM
introduced in VNX
for File 7.0.50.2,
but only 7.0.50.2
or later.

CDES DAE
enclosure
firmware

Common Disk Enclosure Software


(CDES) issue after 42.7 days-will not
recognize new disk, LCC may reset
or possibly hang, causing a DU
situation.

Upgrade to VNX OE for Block


05.31.000.5.007 or later.

ETA
emc265311

Corrected
in VNX for Block
OE 05.31.000.5
.007 and later,
CDES firmware
1.16.

-7-

VNX Issues and Recommended Actions


Issue

Description

Workaround

Solution

Fix

VIA
initialization
fails due to
Cache Dirty
Control LUNs

Cache Dirty LUNs can occur during


system shutdown due to issue with
VNX OE for Block 05.31.000.5.006

If the VIA initialization fails,


output the log from SP A, and
search for Cache Dirty entries.
If "Cache Dirty" entries are
noted, escalate this issue to your
Service Provider.

ETA emc267297.
Escalate all
Cache Dirty LUN
issues to your
Service Provider
for corrective
action.

Contact your
Service Provider
for assistance.

emc261729

Corrected in
VNX for File OE
7.0.35.3.

On the control station:

Corrected in VNX
for Block OE
05.31.000.5.007
and later.

# /nasmcd/sbin/navicli h
<spa_ip> getlog > spa.log
# cat spa.log |grep Cache
Dirty
VIA hangs or
stalls during
setup

VNX Installation Assistant (VIA)


hangs during Pre-Configuration
apply stage

Wait 15 minutes to restart the


VIA application, then use
emc261729 to verify.

-8-

Escalate any
unresolved VIA
or Unisphere
login issues to
your Service
Provider.

Das könnte Ihnen auch gefallen