There are 2 ways,
1. leave the job to run, it will time out as per time out setting
2. If it is less number of target servers, you can bldeploy process on target server.
I do not think so there is any central way to kill specific job.
If you have multiple job servers and can identify which one is currently running this job, you can restart that job server from the Infrastructure Manager in the Console. You would want to confirm no other jobs are running against that same job server before restarting it.
I'm going to bring this one up again because I don't think the previous replies are our only options.
Here's a scenario:
- You kick off three separate CUj's
- Two of the three finish as expected
- The third job is running and according to the task window, it's at 3/3 complete
- You look on the app server the job is running from and you see the job has completed
- You try to kill the job from the Client but does not shut down.
- You try to pause the job and then kill the job but that does not work.
- You wait.
- And wait some more (two days pass).
- You try to kill the job and this still does not work.
Here's the question:
Isn't there a way, from the BladeLogic app server the job is actually running on, to issue a command for just that job to end it?
1 of 1 people found this helpful
Restart the appserver. before you do that though you should look at the ‘appserver status report’ or thread dump and see what the wit for the job is doing and what it’s stuck on. that may also give an idea of how to unstick it other than the appserver restart.
1 of 1 people found this helpful
If there's an external task of any sort that it's hanging on, an NSH process etc., then those can be killed off and will usually free up the thread on the appserver, and allow the job to complete. It's not quite as easy for tasks that run solely within the application server.
Assuming I can find the WIT I guess we can go into the configuration manager component of BladeLogic client and kill it there? Or, is there some sort of nsh command we can run from the app server?
You can’t kill the wit from infra mgr or at all. you can see the stack trace and get an idea of what it’s stuck on – eg doing a nexec or writing to the db.
Oh, btw, one of the responses above says you can run a "bldeploy on target server." What exactly does that mean and are there a bunch of bldeploy commands, with options, one can run? Where can this be found? I've searched the communities and am coming up empty.
not sure what running a bldeploy has to do w/ killing wit...
Nor do I, just felt like asking the question.
We have a similar issue,
Under infrastructure management, a server has two running jobs that are not seen in the tasks in progress screen.
Also navigating through the other job servers I cannot see any associated WIT or lightweight WIT's associated to the job, I'm guessing the only options is a services restart for the job server?
It has also been there since the 21st of April :-/
We're on BSA 8.2.01.
if there are no WIT on any appserver for this job then it's stuck in the database. there's some sql you can run to fix this mostly likely - can you open a ticket w/ support and share the issue # ?
Ahh this makes more sence, as I already have a call open regarding errors in the database cleanup script.
I'm guessing there is a high chance that these two issues are related?
The issue number is ISS04514114
I'm still attempting to find a time where I can run the cleanup script with the servers in debug where users are not affected.
It shouldn’t be related to the cleanup issue.