IV50788: DOC HA 7.1 HOWTO HANDLE SIMULTANEOUSLY REP DISK AND NODE FAILURE

APAR status

• Closed as documentation error.

Error description

• This APAR is applicable for AIX/PowerHA environments where CAA
  (Cluster Aware AIX) clustering is used.
  .
  CAA uses a shared disk across all the nodes in the cluster
  called repository disk. This disk is used for maintaining
  cluster related configuration information. Disk is also used
  for health management through heart beating etc.
  .
  Note that PowerHA and CAA cluster survives loss of repository
  disk. Administrator can recreate the repository disk in such a
  scenario by using the PowerHA SMIT (providing a new disk to be
  used as repository disk)
  .
  This APAR describes an error scenario where loss of repository
  disk occurs and then all the nodes in the cluster also reboot.
  This would result in no configuration information available for
  the nodes to activate the cluster. CAA requires cluster
  configuration on the repository disk to bring up the cluster
  services during boot.
  .
  In this case it is necessary to rebuild the repository disk
  before the cluster services can be activated.
  .
  This APAR provides the PowerHA fixes needed to recover from
  such an error scenario. Note that these PowerHA fixes require
  associated changes in CAA/AIX to operate properly. Please
  obtain the fixes for CAA/AIX using the APARs IV53637 & IV57754
  before proceeding with the PowerHA fix updates..
  .
  Once the CAA/AIX and PowerHA fixes are applied, one can use the
  following process to rebuild the repository disk and then
  activate the cluster services.
  .
  Rebuild Repository disk on one Node
  -----------------------------------
  To rebuild the repository disk execute the following on a node
  (say node A) that has rebooted and also does not have access to
  the original repository disk (Note that cluster will not be
  active due to the unavailability of the original repository
  disk). Note that a new disk will need to be used to rebuild the
  repository using the following process:
  .
  $ smitty sysmirror
  - Problem Determination Tools
  -- Replace the Primary Repository Disk
  .
  Once this step is done, cluster services can be brought online
  on node A.
  .
  Starting Cluster Services on the other nodes of the Cluster
  -----------------------------------------------------------
  Once cluster services have been brought up on node A, other
  nodes in the cluster can be made to join the cluster of node A
  (which is using a new repository disk) using the following
  steps:
  .
  Note that nodes other than node A will continue to access old
  (original) repository disk (if they have access to the original
  repository disk and if the disk is available by now) and would
  need to be pointed to the new repository disk.
  .
  To force these nodes to use the new repository disk, execute
  the following steps on these nodes:
  .
  1.      On each of the nodes in the cluster (nodes other than
  node A) Stop CAA from using the original repository disk
  $ export CAA_FORCE_ENABLED=true
  $ clusterconf -fu
  .
  Check with lscluster -c' that CAA cluster services are inactive.
  .
  2.      From node A, request other nodes to join the CAA
  cluster by executing:
  .
  $ clusterconf ﾖp
  .
  .
  Execute 'lscluster -c' and 'lscluster -m' commands
  to verify that CAA is successfully restarted.
  .
  .
  3.      From Node A Synchronize the new repository disk PowerHA
  configuration information from node A to other nodes.:
  .
  $ smitty sysmirror
  - Cluster Nodes and Networks
  -- Verify and Synchronize Cluster Configuration
  .
  .
  4.      On All nodes other than Node A, start the PowerHA
  cluster Services:
  .
  $ smitty sysmirror
  - System Management (C-SPOC)
  -- PowerHA SystemMirror Services
  --- Start Cluster Services
  

Local fix

Problem summary

• Documentation update for how to handle simultaneously rep disk
  and node failure
  

Problem conclusion

• Documented the steps to handle rep disk and node failure.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

Applicable component levels