Tier1 and Tier-2 Weekly Report 2014-07-29

Tier1 and Tier-2 Weekly Report 2014-07-29

General News

ADC and ATLAS Computing News [20140729]
- ADC Weekly meeting
- JEDI/ BigPandaMon migration
  - Current PandaClient version 0.4.41 default, Jedi version 0.5.5
  - Want to change default very soon (original goal was June/July) - should be transparent to users
  - Confirmed most of the necessary DAST people around (Tadashi, Asoka, Mark)
  - Need BigPandaMon for JEDI support
    - Now have 4 nodes - load balanced serving http://bigpanda.cern.ch/
      - cloud views e.g. http://bigpanda.cern.ch/dash/production/#cloud_CA
      - site views e.g. http://bigpanda.cern.ch/site/CA-SCINET-T2/
    - Started performance benchmarking with pylot and jmeter tools (scaling from 20 to 200 users)
  - Want to announce changes Aug 5th and make change Aug 12
    - Won't really be able to test at scale without real users! Think it is decent time as summer conference results out.
- Rucio commissioning
  - Panda bulk lookup failing
    - “Waiting” jobs
    - Rucio team making some query optimisations
    - Request update of panda servers to latest Rucio client version
  - "Encountered an empty SURL-GUID dictionary"
    - Means failed to contact Rucio for some reason
    - Grid partially drained on Monday due to problem with one Rucio auth server
    - Today another error with as yet unknown cause and solution
  - Rucio-Panda Integration
    - A new DQ2 client allows interaction with both DQ2 and Rucio dataset catalogs
    - Installed on a dev panda instance
    - Plan to also ask HC to use this client
      - Register HC output datasets in Rucio
      - Submit jobs to dev panda instance
      - See what breaks
- Transient data deletion
  - Monthly cleanups have been postponed during migration activity
  - Starting up again - can check datasets slated for deletion at lxplus:~alexei/public/mc*DryRun*
  - If there are no objections these datasets will be removed next Tuesday
- ADCoS report
  - TW recovered from network issues and backup of deletions/srm issues
  - SARA implementing fix for erroneous full dcache tape buffers
  - FZK offline - network hardware dropped GridKA from general purpose internet
  - RRC-KI-T1 transfer failures yesterday, caused by problems with the upstream reverse DNS zone servers - fixed
  - Most sites set to test yesterday early morning, as HC tests were failing, as well as the production jobs. Fixed by rucio devs, as one of the authentication nodes was broken and misconfigured
CA Cloud Squad Actions
- User DaTRI requests approved / processed (last week)
Open Issues and Ongoing Discussions
- Sites should be providing the 2014 storage pledge.
  - REBUS Site Capacities
  - REBUS Federation Resource Pledges
- ATLAS Canada HammerCloud Tests
WLCG and General Computing News
- Security News
- gLite/EMI News
ROC_Canada News
- Current shift