Error code 15230 - Couldn't register the scheduler default with connected server

[tuthmose]

Hi there,

I am trying to install a OpenPBS over RockyLinux. There are two servers and a client (VMs with RL 8.6) and a MoM (physical machine with RL 8.7; I can put 8.6 but not immediatly; installatio of OS wasn’s in my control). I am playing around with options and configuration and thus I restart the scheduling cycle sometimes. Often, but not always, I get this errors:

05/02/2023 11:16:36;0001;pbs_sched;Svr;pbs_sched;Scheduler already connected (15230) in open_server_conns, Couldn't register the scheduler default with connected server
05/02/2023 11:16:43;0001;pbs_sched;Svr;pbs_sched;Invalid request (15004) in open_server_conns, Couldn't register the scheduler default with connected server

which forces me to do a systemctl restart pbs (which would be disruptive in a production environment). I cannot find error code 15230 while 15004 seems a bit generic. This often caused when I enable/disable the cgroups hook following a modification in settings.

[nithinj]

15230: is sched already connected.
Scheduler should not send the registration twice if it can track it. Even if it cannot and decide to send, it should not handle 15230 as an error. Right now scheduler ends up in an infinite loop until it receives a success response from server which it would never get as the scheduler is already registered.
Please file a bug.
you can try killing pbs_sched and start it back rather than restarting all services.

[tuthmose]

Dear nithinj

thank you for your answer; I’ll file a bug. To solve, I kill -HUP pbs_sched and usually it works.