Look in the appserver logs around the time of the scheduled runs – do you see any indication they are trying to run?
Are the jobs scheduled as an active user ?
Thank you for pointing me in the right direction !
Yes, there are messages in the appserver logs:
[Schedule Monitor] Schedule for job 'ASP-Backup N3 - 002' has exceeded the maximum queue time limit of 60 minute(s). It will not be run.
- Yesterday we changed the schedule from 01:00 to 23:00, the job did run at 23:00
- I just found that the jobManagerService has to be restarted after changing MaxJobs
- The service has been restarted and the job is rescheduled for 01:00
I think this will solve the issue, am I correct ?
Is there any information about other parameters that should be changed when MaxJobs is modified from 10 to 20 ?
That message means that it was trying to run for an hour and then gave up before ever starting.
How many physical appservers do you have?
How instances do you have? job/config/etc ?
Max heap per instance?
What version of blade ?
What os are the appservers?
What arch are the appservers?
How many physical appservers do you have? => 1 (vSphere Vrtual Machine, 4 Cores, 8 Gb)
How instances do you have? job/config/etc ? => 1
Max heap per instance? => Maximum JVM Memory (B) = 850,067,456
What version of blade ? => 8.0.817
What os are the appservers? => Windows 2008 R2
What arch are the appservers? amd64
In the console we actually see +/- 23 jobs running and one job with status 'Wating to run ...', the jobs we are running all take some time (> 15 minutes) to complete. This has nothing to do with the performance of the application server, the jobs perform some kind of backup procedure and that just takes some time.
Forgot – how many targets ?
Total number of targets (servers) in Bladelogic is a few hundred, these jobs have only a few targets however.
Think of a batchjob with these steps:
1. Run some powershell scripts against Virtual Center: -> 1 Target : Virtual Center Server
2. Run some powershell scripts against Storage -> 1 Target: Storage Management Server
3. Run some powershell scripts against Virtual Center -> 1 Target: VC
4. -- 6. etcetera
for a x64 system w/ 8g of memory you should be able to set the max heap per instance to 4g - i'm not sure what's it's set to now as that max heap line you sent is allocated, not really the max iirc. the max heap on your instance is seen in the infra structure manager gui if you edit the instance.
it's odd that you have jobs backing up here - is this still happening? are these jobs targeting the same server?
After changing MaxJobs from 20 to 50 I also changed MaxHeapSize, this field was empty in the GUI and I have change it to 2G. The issue isn't happening anymore, however we've seen a few times that jobs were executed while this was not visible : Job->Show Results did not show that the job has run.
During last week everything behaved as it should: all jobs did run fine and the results are visible.
Yes, the jobs are targeting the same server, stupid I didn't realize that this could have consequences. Is there also something like MaxJobsPerServer ?
You may have the jobs queuing up if they are all targeting the same server. or your batch job is set to run in series instead of in parallel.
What is the purpose of these scripts ?