Forum Discussion

backup-botw's avatar
10 years ago

How to kill a hung job that wont cancel

I have run nbrbutil -dump and I can see the below lines in reference to a job I have hung in the queue that just wont cancel out.

         index=6 (Allocation: id={7968DB34-F37A-11E4-8E5C-002128A6F97C} provider=NamedResourceProvider resourcename=xp53tape001.NBU_POLICY.MAXJOBS.Omaha_VMware2 masterserver=xp53tape001 groupid={00000000-0000-0000-0000-000000000000} userSequence=-1 userid="jobid=3036644"  firstuserid="jobid=3036644" named resource allocation)
         index=7 (Allocation: id={7968DB5C-F37A-11E4-AEB0-002128A6F97C} provider=NamedResourceProvider resourcename=xp53tape001.NBU_CLIENT.MAXJOBS.53tape002.botw.ad.bankofthewest.com masterserver=xp53tape001 groupid={00000000-0000-0000-0000-000000000000} userSequence=-2 userid="jobid=3036644"  firstuserid="jobid=3036644" named resource allocation)
         index=8 (Allocation: id={7968D8E6-F37A-11E4-B17F-002128A6F97C} provider=DriveOperationProvider resourcename= masterserver=xp53tape001 groupid={00000000-0000-0000-0000-000000000000} userSequence=0 userid="jobid=3036644"  firstuserid="jobid=3036644" (Media_Drive_Allocation_Record: allocationKey=0 (Media_Drive_Record: MediaKey=0 MediaId= MediaServer=53tape002 DriveKey=0 DriveName= PrimaryPath= PoolName= RobotNum=0 RobotType=0 MediaTypeName= DriveTypeName= NdmpControlHost= RetentionLevel=3 PolicyType=2 JobType=7 MasterServer=xp53tape001) (Storage_Unit_Record: STU= STUType=0 MasterServer=xp53tape001 MediaServer=53tape002 RobotType=0 RobotNumber=0 Density=0 OnDemandOnly=0 ConcurrentJobs=0 ActiveJobs=0 MaxMultiplexing=0 NdmpAttachHost= AbsolutePath=) (Bptm_Strings_Record: ) TpReqFileName=))

Is it possible to do something with nbrbutil -releaseMDS to kill this job? I wasnt sure if I could use the jobid or possibly the allocation ID. Also it looks to have several different allocation ID's so I figured I'd ask.

  • Kill bptm and bpbrm processes on the media server and bpbkar on the client.This should cause the job to exit.
  • You can release the resouces that are used by the job by nbrbutil command . But if you are not able to cancel them, no worries you can getrid of the job once the services of the master is bounced. till then job will be visible in AM. 

  • If the hung job is not getting cleared even after trying to delete. I dont think you will be able to clear it unless services are bounced. 

     

  • This one is my VMware backup host and I cycled services on it completely. Its actually not tied to a specific client...its the parent job for a VM policy.

  • Here the media server processes are still applicable and bpbkar will be running on the backup host (could be media server as well).

  • Yeah when I cycled services on the backup host the one bpbkar that I saw went away and the job is still there. Its not a 50 or a ghost job.

    Looks like I'm stuck having to cycle services on the master. Thank you.