Mailing List Archives
Public Access
|
|
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [HTCondor-users] Replacing my central manager - what interruptions are expected?
- Date: Fri, 18 Dec 2015 15:49:13 -0500
- From: Ben Cotton <ben.cotton@xxxxxxxxxxxxxxxxxx>
- Subject: Re: [HTCondor-users] Replacing my central manager - what interruptions are expected?
Luke,
On Fri, Dec 18, 2015 at 7:20 AM, L Kreczko <L.Kreczko@xxxxxxxxxxxxx> wrote:
> What happens with running/pending jobs when I switch the current
> schedd+manager to be just a schedd and use the new managers?
> Will pending jobs be scheduled on the new cluster or will they be aborted?
> Will running jobs be able to finish gracefully or will they get confused
> (the schedd does not change location)?
>
Jobs that are currently running shouldn't care. Queued jobs will start
running on the new cluster after you've updated the CONDOR_HOST
setting on your scheduler.
> On top of the config change I am also planning to upgrade (at the same time
> or later) HTCondor 8.2.9 to 8.4.3. Same questions apply (are the answers
> listed somewhere in upgrade instructions?).
>
Generally, having mixed versions in a pool isn't a problem (and when
it is, it's almost always a bug). I'm not aware of any
incompatibilities between 8.2 and 8.4. See the HTCondor wiki for
guidance on the graceful upgrade of a pool:
https://htcondor-wiki.cs.wisc.edu/index.cgi/wiki?p=HowToUpgradeCondorGracefully
Thanks,
BC
--
Ben Cotton
Cycle Computing
Better Answers. Faster.
http://www.cyclecomputing.com
twitter: @cyclecomputing