PH17974: AN ACCELERATOR MAY BECOME TEMPORARILY OFFLINE WHILE PREMATURELY SWITCHING INTO SYSTEM STATE 'UNKNOWN'.

APAR status

• Closed as documentation error.

Error description

• An accelerator may become temporarily offline while prematurely
  switching into system state 'UNKNOWN'.
  This issue is accompanied by the following symptoms:
  - The following message will be written into the event log:
    DSNX870I  ... ACCELERATOR <accelerator name> IS NOT ONLINE
    <accelerator name> System went from state 'ONLINE' to
    'UNKNOWN'.
  - Query routing to the accelerator will fail.
  - The accelerator becomes online again after some minutes.
  

Local fix

• For maintenance level 7.1.7 and below, monitor the event log for
  the messages accompanying the issue. If the backend does not
  recover to ONLINE status after 10 or 15 mins, open a Case and
  contact IBM support to check the appliance. Please generate an
  IDAA trace and make sure to have turned on the trace options
  "Accelerator trace information" and "Appliance trace
  information".
  With maintenance level 7.1.8 or higher the temporary connection
  issue that triggered the cluster manager's premature recovery
  actions has been fixed. Consequently, the outage is avoided.
  .
  With maintenance level 7.1.9 or higher the resiliency concerning
  other connection issues between the accelerator and the cluster
  management service will be improved; the timeout for internal
  recovery actions  increased. Consequently, premature reporting
  of an UNKNOWN system state is avoided. 
  

Problem summary

• The issue descibed was fixed in the IDAA V7 accelerator server
  The fix became part of IDAA V7 maintenance level 7.1.9 made
  available on October 24, 2019.
  

Problem conclusion

• Upgrade your IDAA V7 maintenance level to 7.1.9 or higher.
  

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