DUNE / dist-comp

Action items for DUNE distributed computing, and common scripts that are used.
2 stars 0 forks source link

Make Imperial be a rhel8 site #45

Closed StevenCTimm closed 1 year ago

StevenCTimm commented 1 year ago

Hi,

Apologies if this is the wrong list. Also, I'm assuming DUNE runs their workloads in an experiment specific container, if not, you can disregard the rest of this message. We have started upgrading our worker nodes to Rocky 9. Unfortunately the glidein pilots don't run due to an ssl incompatibility. We are aware that DUNE does not support Insert-Your-Favourite- Distribution-Here-9 yet, but we have compat-openssl11 installed, so we think it should work HTCondor EL8, which we think the glide-in WMS can handle. We hope that once the pilots run, the payload will then run in the container. Would it be possible to update the queue behind ceprod03.grid.hep.ph.ic.ac.uk at UKI-LT2-IC-HEP to use EL8 for the pilots ?

Thanks, Daniela

StevenCTimm commented 1 year ago

OSG ticket https://support.opensciencegrid.org/support/tickets/72494 is filed.

StevenCTimm commented 1 year ago

OSG responded, asking DUNE frontend to change to CONDOR_OS = auto We don't want to do that, instead countered asking them to set CONDOR_OS = rhel8 on this entry alone. Waiting for response.

StevenCTimm commented 1 year ago

OSG took our temporary recommendation to set CONDOR_OS for rhel8 for this ceprod03.grid.hep.ph.ic.ac.uk

We will see if it works. Have notified Daniela.

StevenCTimm commented 1 year ago

Daniela reports that it is running

We have to revisit this later when we're ready to run condor 10.2 which has native alma 9 support