I'm having some problems with a scheduled job blocking execution of the others. After logging the scheduler service I've found that the builtin "Subscription" job get stuck in a loop, preventing other jobs on the same site from starting. Here's the events that keep repeating every 10th second:
[2012-04-19 11:20:48UTC] #INF# [_LM_W3SVC_333773749_Root] Execute job 7d8fad39-2682-43b6-b49c-d7c8199d5195[2012-04-19 11:20:48UTC] #INF# 10s warning - due time is den 19 april 2012 09:33:00[2012-04-19 11:20:48UTC] #INF# [_LM_W3SVC_333773749_Root] New scheduled job 2012-04-19 09:33:00 (7d8fad39-2682-43b6-b49c-d7c8199d5195)
09:33 is the time when the first execution was scheduled. The other jobs from the same run fine until this job is firing, jobs on other sites on the same server are not affected.
Any input regarding this would be greatly appreciated!
It seems that the subscription job has got stuck as running. Connect to your database and open the table tblScheduledItem. Look at the column IsRunning if any of the scheduled jobs are marked as running. I guess the subscription job with pkID 7d8fad39-2682-43b6-b49c-d7c8199d5195 is marked as running. If so; set the flag IsRunning to false.
Setting the flag to false gets the jobs up and running again but the problem reoccurs again after a few runs with the "Subscription" job.
Since the returnmessage from last run was ok, I guess the isRunning flag is set to false by the scheduler(?) after the job has executed.
Does anyone know more about this, i.e what can happen, that makes setting the flag fail? I can't find any strange events in the logs around the time of the last run.
Seems to be known bug, have a look http://world.episerver.com/Blogs/Shahid-Nawaz/Dates/2012/1/General-Hotfix-CMS-6-R2/
Luckily, we're upgrading to R2 this week, so it should fix the problem then.