We are facing issue with our tac exec server configuration. We rebuild the servers often like we destroy the tac and execution servers and rebuild everything from scratch except for the Tac DB just to retain the jobs. Now when rebuild we will have two new exec servers and we attach them to tac virtual server.
The issue is when tac comes back it still has the entry on the servers page and the servers are inactive(doesn't exist). Tac keeps polling for the server and eventually we ended up having number of inactive servers on tac which is causing a lot of contention on cpu.
When we try to delete the inactive servers it throws an error stating "Delete failed: Due to related execution task is in progress" below is what is in the error log from tac.
2018-06-08 13:52:25 ERROR TalendRemoteServiceServlet - executionServer.delete.locked
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
Can you please help debug this what could be causing this. this doesnt sound like a bug but something dueto how we rebuild the servers hence i am raising this as discussion.
This is an unusual way of working, but each to his own. Why don't you delete the job servers from the TAC before destroying the servers? I suspect that having the record of the servers in the TAC DB while the servers do not exist, is possibly leading to this. It sounds like when the job server was destroyed it *may* have been either running a job or maybe was scheduled to run a job. However, to get round this you should be able to clear the job server records from the DB if you can find them.
Thanks for your response I will try deleting the server before the rebuild see how it goes. I see this could be an issue due to us not destroying the servers safely.
Having said that if some reason the servers went down unintentionally and I am not able to delete the servers after, is there still a way to delete this. Perhaps delete the entries from the db?