I am not being able to cancel/stop the jobs that have been stuck in launched state.
Can you please tell me ways to do so through Tidal GUI, without disabling the connections and queues.
Not sure what version you are on, I'm on 6.1. Create a job, under Command enter the path of sacmd...
under command parameters enter the command
jobset -i xxxxxx -s 106
Save and submit the job.
I have Tidal installed on Windows so I have a Windows agent where I submit the job using the command line program. I created the job, like any other job, in Client Manager which means is on the TES master.
Oh, I am sorry, I don't use SAP...I use PeopleSoft but I would think the same concept applies no matter where you ran it because you are using the command line program. In my case, the job in TES stays in Launched and does not create a process on PS. I can not just take down the adapter so I had to find another way which is killing the job in TES. I am not sure if you could just add && and add another entry on the same job. Normally I have only 1 job stuck in Launched, out of the blue.
In my experience this behavior is doe the the launcher process failing on the agent. I've only seen it happen on the Linux agents by the way, so maybe this doesn't apply. But, when it happened the only way to cancel the launched jobs is to stop the agent service (it's failed anyway), then cancel the jobs, then start the agent service.
yeah, that what i do now a days, when we are unable to cancel a job stcuck in launched state, we just disable the connection, let the job instance got to orphaned state and then mark it as success.
then enable the connection
Besides getting past the stuck job, is it fixing the issue? When I've seen this happen, it's been an issue on the agent itself and the agent service needed to be restarted. this is different than just disabling the connection.
When I do this, the job is being killed and all I have to do is re-submit it. It's not something that happens often but this way works for me. I am going to try taking down the TES service on the client server maybe every two days and see if this issue still happens. Maybe 6.2 will fix this issue.