Jobs hang in the job controller

Hi,

I have the Mediabin version 4.5.3.

Sometimes there are too many or large assets added into the mediabin will hang the active jobs running in the job controller.
For e.g. I added several assets and one of my active job got hung. The subsequent submitted jobs will also hang in the job controller.
Also sometimes the user uploads a huge file from his laptop, he may disconnect the laptop before the uploading process completes - this will cause the same problem.

When this type of problem occurs, I cancel the hung job but it still does not solve the problem - does not allow the subsequent jobs to run properly. In addition to that, I need to reboot the Mediabin server in order to solve the problem.

How can we solve the issue - to allow the subsequent jobs to run completely and not to be hung in the job controller because of one job hangs?

Joseph Goh

Find more posts tagged with

Comments

Konidala

This could be because of sql server deadlock transactions. Check server log [mbp.log] for deadlocks. Try to Re-index the database which may help to solve the problem. Thanks

https://support.interwoven.com/kb/kb_show_article2.asp?ArticleID=57638

lyman

It is true that SQL deadlock errors appear in the MBP.log frequently in the jobs table. However, if one examines further they are usually benign (although rerunning statistics is always a good idea!).

Generally jobs should clean themselves up. The MediaBin server has a configurable number of slots determined by the setting MaxConcurrentJobs which can be accessed via the MediaBin Enterprise Manager (MeM). You may want to consider increasing the number of these.

However, although really big files can be slow, they should not completely hang. In that case you really should open a support case so we can see what is happening.

Cheers,

Lyman Hurd

rregev

Hi,
This is an issue which we frequently experience however it is usually related to a user uploading a zip file containing a number of files. one job will get stuck in the queue and then all subsequent jobs involving that file type cease to work. I have not yet found out out how\why this happens but a workaround to rebooting your server is to change the HKEY_LOCAL_MACHINE\SOFTWARE\MediaBin,Inc.\MediaBin\Platform\CurrentVersion\CancelRecoveredJobs
registry key from 'never' to 'true'
this will allow you to restart the MediaBin Server service and not have those hung jobs reappear in the job queue. My understanding of what happens behind the scenes is that this registry key will ensure that the Jobs table is cleared when the service is stopped.

regards,
Ronnie Regev
Ubisoft

lyman

To give some background, the server allocates a number of job slots (MaxConcurrentJobs) settable via the MediaBin Enterprise Manager application (in the case of a cluster on a server by server basis). There are circumstances in which jobs "hang" taking down a slot with them (these are rare as almost all job types have a time-out even though in some cases such as MS Office or Multimedia it may be pretty long).

If the server is restarted, the usual assumption is that jobs that had not completed still need to be processed. In most cases this is a reasonable assumption except if the reason for the restart was that a job was getting stuck. This is the purpose of the Never/Once/Always values for the recovery of old jobs.

Many times all that is necessary is to cancel the "stuck" jobs or to increase the MaxConcurrentJobs number for the server freeing another slot. If the problems persist (i.e. they shouldn't!) please do not hesitate to contact support.

Cheers,

Lyman Hurd