OA66487: ENHANCEMENT TO RECOVERY WHEN CONNECTIVITY TO A STRUCTURE FAILS DURING REBUILD

APAR status

• Closed as unreproducible in next release.

Error description

• During a short/temporary loss of connectivity to the only
  available coupling facility by a system (example CF links lost
  to the only available CF), a structure rebuild can be initiated.
  
  If Logger encounters a loss of connectivity during the rebuild
  to the original structure, then logger can process a rebuild
  stop event and issue message IXG107I.
  
  Currently Logger issues ENF 48 signal with
  IXGENFRebuildFailLossConn, IXGENFLogStreamsNotAvailable,
  IXGENFStrRebuildFailed to notify exploiters so that the
  exploiter application can issue IXGCONN REQUEST(DISCONNECT).
  However, logger does not prevent other requests, like CONNECT,
  from taking place.
  
  ANALYSIS:
  Check for bits:
  Strcb_Structure_State = ON
  Strcb_Connectivity_state = OFF
  
  KNOWN IMPACT:
  If an exploiter responds to logger with a different request,
  like CONNECT, Logger allows that request to complete. This can
  lead to IXGRSNCODESTAGINGDSFULL and offloads failing, as
  connectivity to a structure is checked during offload
  processing.
  
  
  ADDITIONAL SYMPTOMS:
  IXC521I REBUILD FOR STRUCTURE RRS_STRUCTURE
  HAS BEEN STOPPED
  
  IXG107I STRUCTURE REBUILD FAILED FOR STRUCTURE RRS_STRUCTURE
  DUE TO A LOSS OF CONNECTIVITY TO THE STRUCTURE.
  

Local fix

• BYPASS/CIRCUMVENTION:
  
  
  RECOVERY ACTION:
  Recycle exploiter
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users of HBB77C0 and above                   *
  ****************************************************************
  * PROBLEM DESCRIPTION: Log stream connector hang (staging      *
  *                      data set full) RC08 RSN865 (IXG071I)    *
  *                      after connector attempts to reconnect   *
  *                      in response to a connectivity loss      *
  *                      during structure rebuild                *
  ****************************************************************
  A log stream connector hang (staging data set full) occurred
  after a structure connectivity loss was detected.  The
  connectivity loss led to the structure being rebuilt.  During
  rebuild, System Logger issues MSGIXG107I and expects the
  connector to issue a DISCONNECT request.
  
  The connector instead performs another CONNECT which it sees
  was successful.  The connector then continues to perform
  WRITEs to the structure.  This eventually leads to the staging
  data set full situation as System Logger does not allow
  offloads to be performed when connectivity loss is detected.
  
  System Logger should not allow CONNECTs, BROWSEs and
  WRITEs when there is connectivity loss to the structure.
  

Problem conclusion

Temporary fix

Comments

• This APAR is being closed UR1 with concurrence from the
  submitting customer. This means that a solution to this APAR
  is expected to be delivered from IBM in a release (if any)
  to be available within the next 36 months.
  

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

  SYSTEM LOGGER

• Fixed component ID

  5752SCLOG

Applicable component levels