IV95025: CAA DOES NOT START AFTER REPDISK REPLACEMENT WITH SAVEBASE ERRORAPPLIES TO AIX 7100-04

APAR status

• Closed as program error.

Error description

• When a repository gets replaced the clvdisk
  attribute of cluster0 device in ODM is updated
  with the UUID of the new primary repository disk.
  The savebase command is executed to update the
  Boot Logical Volume.
  
  In case an entire data centre fails:
  * One of two PowerHA/CAA nodes fails
  * One of two storage subsystems fails.
    The failed storage subsystem hosts
    - the primary repository disk
    - the rootvg disk of the remaining node,
      which has been used at boot.
  
  the savebase of the Automatic Reposirory
  Replacement (ARR) operation fails.
  syslog.caa looks like:
  ...
  Apr  3 12:36:54 ha1clC caa:info cluster[8454150]:
   cluster_utils.c        cl_run_log_method
    11951   523    START '/usr/sbin/chdev -l cluster0
     -a clvdisk='2e199ead-1ed0-aeb4-e331-8c59c3f668f6''
  ...
  Apr  3 12:36:55 ha1clC caa:info cluster[8454150]:
   cluster_utils.c         cl_run_log_method
     11951   523   START '/usr/sbin/savebase '
  Apr  3 12:36:55 ha1clC caa:info cluster[8454150]:
   cluster_utils.c         cl_run_log_method
    11982   523     FINISH return = 1
  ...
  Apr  3 12:36:55 ha1clC caa:err|error cluster[8454150]:
   caa_config.c     cl_th_cfg_msg   6691    523
     savebase failed on ha1clC.mainz.de.ibm.com.
     Please run savebase on ha1clC.mainz.de.ibm.com.
  ...
  Apr  3 12:36:57 ha1clC caa:info cluster[8454150]:
   cl_chrepos.c  automatic_repository_update
    2297    1       FINISH rc = 0
  Apr  3 12:36:57 ha1clC caa:info cluster[8454150]:
   caa_protocols.c       recv_protocol_slave
    1542    1       Returning from Automatic
     Repository replacement rc = 0
  ...
  
  CAA will not start after a reboot of
  remaining node. syslog.caa looks like:
  ...
  Apr  3 14:02:37 ha1clC caa:err|error cluster[3473558]:
   cluster_utils.c  cluster_repository_read_data
    4777    1       Could not get name of cluster
     repository disk from ODM (ODMDIR=/etc/objrepos).
  Apr  3 14:02:37 ha1clC caa:info cluster[3473558]:
   cluster_utils.c       cl_kern_repos_check
    11858   1       Could not read the respository.
  ...
  Apr  3 14:02:37 ha1clC caa:err|error
   cluster[3473558]:   clusterconf_lib.c
     _find_and_load_repos     1482    1
      cluster_repository_query() found a UUID
       but no corresponding disk. This condition
       may be temporary.
  Apr  3 14:02:37 ha1clC caa:warn|warning
   cluster[3473558]:    1035-242 clusterconf:
    Non-fatal error when loading the topology.
  ...
  

Local fix

• Run
  $ clusterconf -r <new primary rep disk>
  at the remaining node after reboot.
  

Problem summary

• savebase can fail on a mirrored rootvg if one disk
  is not accessible.
  

Problem conclusion

• made savebase to look for blvs on all the disks in the mirrored
  rootvg.
  

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:

  IV97722 IV97730 IV97733 IV97748 IV97751 IJ01657 U882085

Fix information

• Fixed component name

  AIX V7.1

• Fixed component ID

  5765H4000

Applicable component levels

• R710 PSY U882085

     UP18/08/22 I 1000