After performing a migration or version upgrade in IBM AIX PowerHA (formerly HACMP) environments, administrators may encounter cluster version mismatches.
This typically occurs when the Object Data Manager (ODM) retains old version information, leading to inconsistencies that prevent PowerHA services from starting or functioning correctly.
This guide explains how to safely identify and correct PowerHA ODM version mismatches, ensuring cluster integrity and successful startup.
Step 1: Backup ODM (On Both Nodes)
Before making any changes, always back up the ODM repositories on both cluster nodes to prevent data loss in case of an error.
# tar -cvf hacmp.odm.tar /usr/es/sbin/cluster/etc/objrepos
# tar -cvf system.odm.tar /etc/objrepos
Tip: Keep these backups in a safe location. They can be restored if any configuration corruption occurs.
Step 2: Correct the HACMP/PowerHA Cluster Version (On Both Nodes)
Export the Current HACMPcluster Object Class
Dump the current cluster object data into a file:
# odmget HACMPcluster > cluster.file
Review the File
Example output:
# cat cluster.file
PowerHA SystemMirror cluster:
id = 1106245917
name = "HA72_TestCluster"
nodename = "node1"
sec_level = "Standard"
sec_level_msg = ""
sec_encryption = ""
sec_persistent = ""
last_node_ids = ""
highest_node_id = 0
last_network_ids = ""
highest_network_id = 0
last_site_ids = ""
highest_site_id = 0
handle = 1
cluster_version = 17
reserved1 = 0
reserved2 = 0
wlm_subdir = ""
settling_time = 0
rg_distribution_policy = "node"
noautoverification = 0
clvernodename = ""
clverhour = 0
3. Edit the Cluster Version
Open the file in a text editor:
# vi cluster.file
Locate the line:
cluster_version = 17
Update the version number to match your installed PowerHA version.
For example:
PowerHA/HACMP Version cluster_version Value
HACMP 7.2 17
PowerHA 7.3 20
Make sure that the cluster name and nodename fields are correct and consistent across nodes.
4. Remove the Existing ODM Object
Delete the old cluster object from ODM:
# odmdelete -o HACMPcluster
5. Re-add the Corrected ODM Object
Add the updated configuration back into the ODM:
# odmadd cluster.file
6. Verify the Changes
Confirm that the cluster version and attributes were updated successfully:
# odmget HACMPcluster
Ensure the cluster_version field now reflects the correct version number.
Verify and Synchronize Cluster Configuration (Primary Node Only)
Once the ODM has been corrected on both nodes, perform a cluster verification and synchronization from the primary node (NodeA):
# smitty hacmp
Navigate through the SMIT menu:
Cluster Applications and Resources → Verify and Synchronize Cluster Configuration
This process ensures that all nodes have identical and valid cluster configurations.
Start PowerHA Cluster Services (Primary Node Only)
After successful verification and synchronization, start the cluster services from NodeA:
# smitty hacmp
Navigate:
System Management (C-SPOC) → Manage HACMP Services → Start Cluster Services
Select Both Nodes (NodeA and NodeB) when prompted.
Step 7: Verify Cluster and Resource Status
Once the cluster is running, check that all resources are online and functioning correctly.
# clstat
# clshowsrv
# lsvg -o
# df -g
# ifconfig -a
Confirm that:
- Volume groups are varied on,
- Filesystems are mounted,
- Service IPs are active, and
- Applications are running as expected.
No comments:
Post a Comment