Batch Operations Weekly

31/S-023 (CERN)



Show room on map
Videoconference Rooms
Gavin McCance
Auto-join URL
Useful links
Phone numbers
    • 14:00 15:00
      Agenda 1h
      • Misc:
        • INC1995937: VMs stuck in Geneva Project 004 (CMS T0) & Geneva Project 008 (AMS T0). (FIXED)
        • OS-9136 OS-9182: Issues with Neutron in the batch OpenStack region due to scalability issues: < 50% success rate.
      • Haggis (API wrapper):
      • Consul & friends:
        • Two deployment models being tested: puppet & k8s.
        • Provide a production like env: encryption, ACLs, TLS,…
        • First use-case: terraform state.
        • Testing a new simplified rundeck deployment (and upgrading to 3.x).
      • New EPIC Batch on Azure: BBC-2329
        • Cost management permissions fixed in IT Batch subscription.
        • Reader application credentials created. This allows to query from grafana (it requires Grafan 6.x, only monit-grafana-dev at the moment).
        • Next: test all the budget/cost mechanics to control resource groups.
      • Datacentre fact for point 8, decide: [BBC-2333] - Project
      • Point 8 to be drained for Monday for IT-CS
        • Will not be doing cloud region at same time
      • Need to deco / recreate Wigner CEs [BBC-2334]
        • recreate with same name so downtime only
      • INC1998136 One argus backend had a bad certificate, CRAB issues (though for longer than should be)
      • gfal2 packages now in cc7
      • BEER security: decision from MM at this point is that we continue to take production jobs, and IT-ST patch at their convenience
      • EOS FUSE: development team confident bugs we’d highlighted have been fixed. Test plan from now:
        • More ad-hoc testing by us
        • Kuba has several accounts he can now use to submit from EOS.
        • Testing by ST ops & us for a couple of weeks, then will find a friendly use case (AMS possibly)
      • Next week: barbie