2683098 - Multiplex coordinator node crash during restart at call db_Catalog::GetLocalWritableLock - SAP IQ

• On restart of the multiplex coordinator node the server crashes at st_server.cxx with this stacktrace
  
  

  db_Catalog::GetLocalWritableLock
  db_TableVersion*db_Catalog::FindTableByID
  db_Catalog::GlobalRICheck
  db_GlobalRICheck::Execute()
  st_command::DoCmdThroughResourceGate()
  st_iqdml::GlobalRICheck
  UIQDML_GlobalRICheck
  st_SAIQInterfaceInfo::callFunction()
  st_SAIQInterface::RunIQFunc
  st_SAIQInterface::Execute
  saint_iqthresholddml::GlobalRICheck
  Database::RestoreGlobalTxnContext
  
  

• The crash has consistently been seen during Global Transaction Restore recovery processing. The problem happens after seeing this message in the iqmsg file... 

st_database::CompleteMpxRecovery() - change RecoveryState from RECOVERED to RECOVERING 

  ...and before the expected recovery completion message that follows.  The crash happens and this message is not printed.

GBL_TXN_RES Global Transaction Restore time : n.nnnn

• The problem requires that there be at least one active secondary writer node in the mpx environment and that global transaction resiliency functionality is enabled.
  
  
• In some cases if the writer node is waiting for completion of pending DML transactions the stack trace will contain a call to st_txncb::GetEffectiveVersioningType
  
  

• multiplex environments only
  
  
• SAP IQ 16.1 SP02 all platforms
  
  
• SAP IQ 16.1 SP03 all platforms

sybase, co-ordinator, abort, fail, CR815362, 815362, CR#815362, CR815141, 815141, CR#815141, GTR, restore, hang, crashed, 'signal 11', sig11, 'stack trace', stktrc, recovery, failure, 'thread stacks', stcxtlib , KBA , BC-SYB-IQ , Sybase IQ , Problem