IV97265: CAA:SLOW GOSSIP TRANSMISSION ON BOOT MAY CAUSE PARTITIONED CLUSTAPPLIES TO AIX 7100-04 17/09/29 PTF PECHANGE

APAR status

• Closed as program error.

Error description

• **************************************************************
  * USERS AFFECTED:
  * Systems running the AIX 7100-04 Technology Level
  * with bos.cluster.rte at the 7.1.4.30 or 7.1.4.31 level.
    **************************************************************
  * ERROR DESCRIPTION:
  * After rebooting a node in either a PowerHA or VIOS SSP cluster
  * using CAA, there is a chance that the node may create its own
  * cluster, causing a split-brain / partitioned cluster in the
  * CAA environment.
  *
  * This is more likely to be seen if the network is slow and
  * there is a delay in gossip packets being received by the
  * rebooted node.
  *
  * The effect of a split-brain / partitioned cluster can vary,
  * but in the worst cases: PowerHA may react by bringing
  * resources online at the same time on multiple nodes, and
  * VIOS SSP can experience pool going down on one or more nodes.
    **************************************************************
  * RECOMMENDATION:
  * Install APAR IV97265.
  * Prior to fix availability, an interim fix is available from
  * either
  * ftp://aix.software.ibm.com/aix/ifixes/iv97265/
  * https://aix.software.ibm.com/aix/ifixes/iv97265/
  * Installation of the ifix requires a reboot.
    **************************************************************
  

Local fix

• n/a
  

Problem summary

• PROBLEM SUMMARY:
  After rebooting a node in either a PowerHA or VIOS SSP
  cluster using CAA, there is a chance that the node may
  create its own cluster, causing a split-brain / partitioned
  cluster in the CAA environment.
  This is more likely to be seen if the network is slow and
  there is a delay in gossip packets being received by the
  rebooted node.
  The effect of a split-brain / partitioned cluster can vary,
  but in the worst cases: PowerHA may react by bringing
  resources online at the same time on multiple nodes, and
  VIOS SSP can experience pool going down on one or more
  nodes.
  

Problem conclusion

• There is a gate in which all initial clusterwide lock
  requests should consider the count of nodes heartbeating to
  the repository in addition to those gossiping over network.
  There was a hole in the gate and the fix closes it.
  

Temporary fix

•   *********
    * HIPER *
    *********
  

Comments

APAR Information

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

  IV97148

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

Fix information

• Fixed component name

  AIX V7.1

• Fixed component ID

  5765H4000

Applicable component levels

• R710 PSY U873626

     UP17/09/22 I 1000