Beruflich Dokumente
Kultur Dokumente
IBM
pSeries
server
IBM
Shared disk
Critical applications
Database server
Web server
Dependent on DB
pSeries
serve
r
IBM
Web Srv
Private
Network
pSeries
serve
r
B
Shared Disk
2008 IBM Corporation
pSeries
serve
r
IBM
Web Srv
Private
Network
pSeries
serve
r
B
Shared Disk
pSeries
serve
r
IBM
Web Srv
Private
Network
pSeries
serve
r
B
Shared Disk
Failover possibilities
One to any
One to one
I
Any to one
Any to any
Fallover Preferences
Fallback Preferences
Fallback To Higher Priority Node - (FBHP)
Never Fallback - (NFB)
7
Common Resources
Service IP Address
Highly available address(es) that users/apps/clients use for connectivity.
Application
Highly available application that PowerHA controls via:
Application Start Scripts
I
Application Stop Scripts
Application Monitoring
Volume Groups
Shared Data Volume Groups
Can be raw logical volumes or Filesystems
Defaults to mount all filesystems (if applicable)
10
11
12
Online Planning
Worksheets For
Resource Groups
Shown Here
13
Hardware
Speed
of Takeover
Firewall
Issues
IPAT via Replacement
Node A
net_ether_0
Node B
en0 9.19.10.1
(boot)
en0 - 9.19.10.28
(service IP)
14
Speed
of swap
Hardware
Firewall
net_ether_0
Address Takeover
(HWAT)
I
Issues
Node A
Node B
9.19.10.28
(persistent a)
9.19.10.51
(service IP)
9.19.10.29 (persistent b)
en1 192.168.11.2 (base2)
(service IP)
2008 IBM Corporation
Network Switches
Node A
ent0
ent1
ent2
en2
Network: net_ether_0
192.168.10.1 base
192.168.10.1 persistent a
Node B
base 192.168.10.2 en2
ent2
ent0
ent1
persistent b 192.168.10.1
9.19.10.50 service IP
Verification Messages:
For nodes with a single Network Interface Card per logical network configured, it is
recommended to include the file '/usr/es/sbin/cluster/netmon.cf' with a "pingable IP
address as described in the 'PowerHA Planning Guide'.
WARNING: File 'netmon.cf' is missing or empty on the following nodes: glvm1 glvm2
Tip:
* Single adapter networks can be just as redundant and require less IPs
16
Examples - Collocation
en0
en0
en0
en1
en1
en1
en1
en2
en2
en2
en2
1500
1500
1500
1500
1500
1500
1500
1500
1500
1500
1500
link#2
192.9.201
192.9.168
link#3
192.9.201.1
1.1.1.1
1.1.1.2
link#4
192.9.202
1.1.1.1
1.1.1.2
0.6.29.dc.82.ca
ppstest3_enboot
ppstest3_persist
0.6.29.b9.1f.71
ppstest3_enstby1
service_1
service_2
I
0.6.29.dc.82.86
ppstest3_enstby2
service_1
service_2
BOOT
PERSISTENT
BOOT
SERVICE
SERVICE
BOOT
SERVICE
SERVICE
Policy does not account for the location of the persistent label - it is possible
for all addresses to be on one interface.
Subsequent failures may cause the service labels to move to a different
interface, but they will all be mapped to the same interface.
18
1500
1500
1500
1500
1500
1500
1500
1500
1500
1500
link#2
192.9.201
192.9.168
1.1.1.1
1.1.1.2
link#3
192.9.201.1
192.9.168
1.1.1.1
1.1.1.2
0.6.29.dc.82.ca
ppstest3_enboot
ppstest3_persist
service_1
service_2
0.6.29.b9.1f.71
ppstest3_enstby1
I
ppstest3_persist
service_1
service_2
BOOT
PERSISTENT
SERVICE
SERVICE
BOOT
PERSISTENT
SERVICE
SERVICE
1500
1500
1500
1500
1500
1500
1500
1500
1500
1500
1500
1500
link#2
192.9.201
192.9.168
link#3
192.9.201.1
1.1.1.1
1.1.1.2
link#4
192.9.202
192.9.168
1.1.1.1
1.1.1.2
0.6.29.dc.82.ca
ppstest3_enboot
ppstest3_persist
0.6.29.b9.1f.71
ppstest3_enstby1
service_1
service_2
0.6.29.dc.82.86
ppstest3_enstby2
I
ppstest3_persist
service_1
service_2
BOOT
PERSISTENT
BOOT
SERVICE
SERVICE
BOOT
PERSISTENT
SERVICE
SERVICE
Configuration
Extended Configuration -> Extended Resource Configuration
HACMP Extended Resources Configuration
Configure Resource Distribution Preferences
Configure Service IP Labels/Address Distribution Preference
Select the Network to Change Service Label Distribution Preference
Move cursor to desired item and press Enter.
net_ether_01
[Entry Fields]
net_ether_01
Anti-Collocation
+
2008 IBM Corporation
Remove
Exportvg
Show
autocorrective actions
Review
22
cluster configuration
Reasons to use:
Some models have serial port restrictions for heartbeating on integrated ports
Clusters with more than two nodes may require an async adapter with a RAND
Requires:
Installation of bos.clvm.enh
Best Practice:
Although more redundant networks are better if you had to choose - disk
heartbeating is tried and tested and should replace if not coexist with RS232
networks as long as there is a shared SAN infrastructure in place.
23
[Entry Fields]
jordan_rg
jordan jessica
Online On Home Node O>
Fallover To Next Prio>
Never Fallback
[loki_svc1]
[]
+
+
Volume Groups
Use forced varyon of volume groups, if necessary
Automatically Import Volume Groups
Filesystems (empty is ALL for VGs specified)
Filesystems Consistency Check
Filesystems Recovery Method
Filesystems mounted before IP configured
Filesystems/Directories to Export
[]
false
false
[]
fsck
sequential
false
[]
[]
[]
+
+
+
+
+
+
+
+
+
Tape Resources
Raw Disk PVIDs
[]
[]
24
+
+
+
2008 IBM Corporation
25
I
Guaranteed system resources
for multiple applications
Parent/Child Relationships
Great for Multi-Tier environments
26
RG A
RG B
RG A
RG C
RG B
RG C
RG D
27
28
29
Non-concurrent
rgDB has High priority, rgApp has Intermediate priority, rgWeb and rgTest
I
have Low priority
nodeA
nodeB
nodeC
nodeD
Production Server
Production Server
Production Server
Standby Server
rgDB
High Priority
rgApp
Intermediate Priority
30
rgWeb
rgTest
Low Priority
Low Priority
nodeA
nodeB
I nodeC
nodeD
Production Server
Production Server
Production Server
Standby Server
rgDB
rgDB
High Priority
High Priority
rgApp
Intermediate Priority
31
rgWeb
rgTest
Low Priority
Low Priority
rgTest
OFFLINE
due to lack
of node
nodeA
nodeB
Production Server
Production Server
nodeC
nodeD
Production Server
Standby Server
rgDB
High Priority
32
rgApp
rgApp
Intermediate Priority
Intermediate Priority
rgWeb
rgWeb
rgTest
Low Priority
Low Priority
Low Priority
SMIT Configuration
smit cm_rg_dependencies_menu
->Extended Configuration
->Extended Resource Configuration
->Configure Resource Group Run-Time Policies
->Configure Dependencies
between Resource Groups
I
Configure Dependencies between Resource Groups
Move cursor to desired item and press Enter.
Configure
Configure
Configure
Configure
33
Parent/Child Dependency
Online on the same node Dependency
Online on different nodes Dependency
Online on the same site Dependency
DLPAR/CUoD configuration
PowerHA on the primary machine detects the failure
Running in a partition on another server, PowerHA grows the backup
partition, activates the required inactive processors and restarts
application
DLPAR/CUoD Server
(running applications on active processors)
I
Production
Database Server
PowerHA
HACMP
34
Shared
Disk
Inactive Processors
Order Entry
Database
Server
Web Server
Active Processors
PowerHA
HACMP
2008 IBM Corporation
Application Monitoring
PowerHA can monitor applications in one of two ways:
I
Restart Can establish a number of restarts to restart locally. After a
specified restart count, if app continues to fail you can escalate to a
fallover.
Notifiy Send email notificaiton
Fallover Move application and associated resource group to next
candidate node.
35
Application Monitoring
Add a Custom Application Monitor
Type or select values in entry fields.
Press Enter AFTER making all desired changes.
*
*
*
*
Monitor Name
Application Server(s) to Monitor
Monitor Mode
Monitor Method
Monitor Interval
Hung Monitor Signal
* Stabilization Interval
* Restart Count
Restart Interval
* Action on Application Failure
Notify Method
Cleanup Method
Restart Method
36
[Entry Fields]
[testmon]
+
[Long-running monitori> +
[/usr/lpp/custom/testapp]
[300]
#
[]
#
[600]
#
[3]
#
[600]
#
[notify]
+
[]
[/usr/lpp/custom/appclean]
[/usr/lpp/custom/appstart]
37
Configuration_Files
HACMP_Files
2008 IBM Corporation
Automatic Propagation:
39
40
41
Connect
Start
Cluster services
Show
42
to cluster
43
44
45
46
47
49
50
A Cluster Test Tool Test Plan is a flat text file with one line for each desired test.
Each line in the test plan specifies a test to be run and the parameters that are
required.
The last field on each line is an optional test description. It will be included in the
output and the log file if it is present.
The tests will be executed in the order in which they appear in the test plan
If the first character on the line is a '#', the line is a comment and will be ignored.
Blank lines are also ignored.
51
PowerHA Administration
Most common tasks performed via C-SPOC
I
Most common task include
52
PowerHA Administration
C-SPOC Menu (smitty cl_admin)
System Management (C-SPOC)
F2=Refresh
F10=Exit
F3=Cancel
Enter=Do
F8=Image
54
Trouble Shooting
Generally involves:
55
56
Trouble Shooting
Determining problem source:
Both PowerHA and AIX offer many diagnostic tools. Common
PowerHA tools include:
clstat
cldump
clRGinfo
Trouble Shooting
From a support perspective it is common to start by reviewing the logs in
search of the error cause. The following is the most common logs used.
Complete list can be found in Chapter 2 of the HACMP Trouble Shooting
Guide.
Cluster log files:
58
60
Gateway design
Enterprise
clusters.
POWER 6 DLPAR Support
IPv6 Support
First version to utilize non-disruptive upgrades
Smart-Assists updated for currency
61
Monitor
62
WebSMIT status
I
Show
Show
Thank You
I
Questions?????
63