MVSFORUMS.com

gcs · Beginner Joined: 28 Aug 2006 Posts: 3 Topics: 1

We are trying to run a BMP DL/I program that has got IMS calls only. In the program we are taking a Checkpoint using CHKP call. After abending the program, trying to restart using XRST call, then the program is abending with following error:

+DFS1000I IMS910 ABEND U0102 MODULE=DFSZSR00 LABEL=E2D90004 R14=800344B0 DLA=C4D3A002 IMS#
DMB676I SUDUMPM DUMPMASTER V6.100P DMBINTRH CONTINUING WITH DUMP-TAKING FOR JOB #######
DMB615I SUDUMPM V6.100P DMBINTRH STARTING TO TAKE DUMP NO 460 FOR JOB #######
DMB170I SUDUMPM V6.100P DUMP NO. 460 RECORDED IN BDF -
DMB616I SUDUMPM V6.100P DMBINTRH DUMP NO 460 COMPLETED
DMB604I DMSCSVCR PASSING DUMP TO IBM
IEA995I SYMPTOM DUMP OUTPUT 411

The program originally abended because of a IMS status of XD after an IMS checkpoint was performed.

XD Explanation: IMS is terminating by a CHECKPOINT FREEZE or DUMPQ. IMS returns this code to a BMP that has issued a CHKP or SYNC call. If it is a transaction-oriented BMP, IMS does not return a message.

Any help would be greatly appreciated!

- GCS

Bithead · Posted: Mon Aug 28, 2006 8:45 am Post subject:

The original abend looks like IMS was brought down while the jobs was running.

The restart is failing with a U0102 - a positioning error. Make sure that any GSAM files that you are using have not changed since the program took its last checkpoint.

gcs · Beginner Joined: 28 Aug 2006 Posts: 3 Topics: 1

Bithead · Posted: Mon Aug 28, 2006 9:31 am Post subject:

Were the databases touched? Did another run against them? Also check that the online log is available.

gcs · Beginner Joined: 28 Aug 2006 Posts: 3 Topics: 1

No the databases were not touched. No other program ran against them. Yes, the online log was available.

It looks like we used the key work LAST for the IMS program restart. It looks like the IMS checkpoint id had been dropped when the utitlity tried to replace LAST with the correct checkpoint id.

All we need to do was look a the program output and retrieve the last checkpoint id and hardcode it into the restart field.

One last question:
--------------------------------------------------------------------------------------------
Background:
When looking at the 1st abend output I see the last checkpoint list as:
DFS0540I *EXTENDED CHECKPOINT* PROGRAM=PROGRAM1,CKPTID=00000389,REGID=0055,DAY=239,TIME=0602006 IMS#
DFS0542I *EXTENDED CHECKPOINT* DAY=239,TIME=060212085206-0500 IMS#

This is the checkpoint call that returned the XD status code.

Question:
Should I use the last checkpoint id above (00000389) or the one right before it (00000388) to restart the program correctly?

- GCS

Bithead · Posted: Mon Aug 28, 2006 12:20 pm Post subject:

You should use the last checkpoint from the run that originally failed when IMS was brought down.

Make sure that the online log has not been archived. If it has, code in an IMSLOGR referencing the log archive dataset name.