Task stalled waiting for ICEECB when an Archive Journal took an extra long time to condense.

Document ID : KB000014126
Last Modified Date : 14/02/2018
Show Technical Document Details
Question:

WE had the following messages and a task Stall abend when an Archive Journal condense took longer than usual.  What does this mean and what is ICEECB? 

23.58.31 STC00722 +IDMS DC203004 Vnnn Tnnnnnnn SYSJRN08 Being Condensed - Waiting for Finish 
23.58.40 STC00722 +IDMS DC203004 Vnnn Tnnnnnnn SYSJRN08 Being Condensed - Waiting for Finish 
23.58.50 STC00722 +IDMS DC203004 Vnnn Tnnnnnnn SYSJRN08 Being Condensed - Waiting for Finish 
23.59.00 STC00722 +IDMS DC001003 Vnnn Tnnnnnnn TASK:BATCBULK PROG:<pgmname> STALLED WAITING FOR ICEECB AT 2FF83C9C 
23.59.00 STC00722 +IDMS DC201006 Vnnn T1 ABRT D002 nnnnnnn 100 BATCBULK xxxxxxx xxxxxxxxx xxxxxx nnnnnnnn 
23.59.13 STC00722 +IDMS DC203004 Vnnn Tnnnnnnn SYSJRN08 Being Condensed - Waiting for Finish 
23.59.22 STC00722 +IDMS DC203004 Vnnn Tnnnnnnn SYSJRN08 Being Condensed - Waiting for Finish 
00.01.12 STC00722 +IDMS DC203005 Vnnn Tnnnnnnn Program-ID <pgmname> Transaction-ID nnnnnnnnnn has been Rolled Out! 

Answer:

In this case a task needed access to the Disk Journal in question for automatic recovery of the transaction.

The Disk Journal in question was in the middle of the Condense phase of ARCHIVE JOURNAL so the online transaction has to wait until it completes.

CV will put out the DC203004 message and then put the task into a WAIT for 10 seconds (using #SETIME TYPE=WAIT so the wait is for ICEECB) and when the timer pops it will check the Journal to see if the recovery can now proceed.

This is a rare occurrence but client had INACTIVE INTERVAL set to 10 seconds as well, so one of those 10 second intervals, the Stall interval (INACTIVE INTERVAL) popped before the ICEECB timer popped.

Again this is rare but because we have a set interval of 10 seconds for rechecking the journal, we suggested they set INACTIVE INTERVAL to 11 seconds.