Batch queue not processing jobs; aleph_shutdown stalls shutting down batch que

• Article Type: General
• Product: Aleph
• Product Version: 16.02

Description:
Various jobs have been submitted to the xxx50 batch queue, both from the job_list and from the GUI Services, in recent days, but none of these jobs have been executed: there is no output for them in xxx50 $data_print and there is no log in $alephe_scratch.

We have have stopped/restarted the batch que (util c/3, util c/2) several times, but that has not helped.

Also, we find that the aleph_shutdown stalls in trying to stop this batch queue.

Resolution:
This can happen when the batch que (lib_batch) process is not stopped by aleph_shutdown prior to a shutdown of Oracle.

Do the following:

1. util c/1: check the date of the process; if it is prior to the last time you did the util c/3 "stop", the stop has not actually been working.

2. util c/7 to see if any jobs show in the queue; if not, then go to step 11.
If the jobs which show are all ones which can be resubmitted without any problem, go to step 11.

3. df1, then "vi que_batch".
4. util c/1; note the process number nnnn (the first number) of the lib_batch process
5. kill -9 nnnn <where nnnn is the number from step 4>
6. util c/2 to start the batch queue (this is necessary to make the jobs viewable)
7. util c/3 to stop the batch queue {this lets you delete duplicates (step 9) before the jobs actually run}
8. util c/7 -- you will now see all the jobs which have been waiting to run
9. If you see duplicates and want to delete them, do util c/8 to delete entries from the queue
10. util c/2 to restart the batch queue

If the preceding does not work (that is, the jobs still do not run), then proceed to step 11

11. df1, "rm que_batch" (to delete the existing one), then "touch que_batch" (to create a new, empty one)

You should now find that util c/3 successfully stops the queue (and the shutdown doesn't stall at the step for this library).

See also KB 5672.

• Article last edited: 10/8/2013