PI56154: After a network bronwn out recovers, the routetable contains unr eachable entries

APAR status

• Closed as program error.

Error description

• After a network brownout recovers, the routetable contains
  unreachable or entries pointing. The routetable entries that are
  unreachable match the locations of shards on servers that were
  stopped or restarted during brownout recovery.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:  WebSphere eXtreme Scale or WebSphere        *
  *                  DataPower                                   *
  *                  XC10 Appliance users who experience a       *
  *                  network                                     *
  *                  brownout (a temporary network failure       *
  *                  lasting                                     *
  *                  longer than 30 seconds or the configured    *
  *                  failover detection time).                   *
  ****************************************************************
  * PROBLEM DESCRIPTION: After a network bronwn out recovers,    *
  *                      the                                     *
  *                      routetable contains unreachable         *
  *                      entries                                 *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  After a network brownout recovers, the routetable contains
  unreachable or entries pointing. The routetable entries that
  are
  unreachable match the locations of shards on servers that were
  stopped or restarted during brownout recovery.
  This can happen if a catalog server is isolated with a group of
  containers and does some shard movement while islanded,
  incremented the placement work ID for their grids. If the group
  rejoins the domain or collective, the containers can push their
  route table entries to the catalog servers. If the same
  containers are then restarted to recover from the brownout, then
  bad route table entries can be left in the route table and new
  routes temporarily rejected because they appear to be stale or
  old.
  

Problem conclusion

• The route table processing double checks if potentially stale
  route entries are actually valid (match the current placement
  plan). This allows "bad" routes from a network brownout failure
  be overridden.
  

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

• Fixed component name

  WS EXTREME SCAL

• Fixed component ID

  5724X6702

Applicable component levels

• R860 PSY

     UP