[HTCondor-users] restarting / reloading / re-config

[Jeff Templon <templon@xxxxxxxxx>]

Hi Folks,

Weâre bringing HTCondor under our config mgt system.  We have the 
following question about how to deal with a changed config. In the past, 
on non-HTCondor machines we usually restart the daemon.  We have seen 
that restarting daemons on HTCondor generally results in a restart of 
jobs (unfortunate, this is [ use your best Yoda accent here ]).  
condor_reconfig works, but does a reload work as well?

Forwarded message:

> From: Dennis van Dok <dennisvd@xxxxxxxxx>
> To: Jeff Templon <templon@xxxxxxxxx>
> Subject: Re: ganymede nodes in pre-prod
> Date: Fri, 19 Mar 2021 15:04:01 +0100
>
> On 18-03-2021 20:24, Jeff Templon wrote:
> >  Reason: a restart of the condor daemons will restart all the jobs. 
> > Can salt handle this, that on condor nodes when a config change is 
> > made, it runs `condor_reconfig` instead of restarting the condor 
> > daemon via `systemd`??
>
> Would a systemctl reload condor do the same thing? It seems to send 
> SIGHUP to the master. That would be the easiest solution.

JT