Tier1 and Tier-2 Weekly Report 2014-04-15

Tier1 and Tier-2 Weekly Report 2014-04-15

General News

ADC and ATLAS Computing News [20140415]
- ADC Weekly meeting
- Rucio Migration
  - FZK today
  - Already done for DE T2s
  - UK Cloud next week
  - Some problems seen with network connections from WNs - add a "rucio ping" to pilot wrappers
  - Some futher cleanup of AGIS paths
  - Reaper daemon switched on this week for test datasets first
  - Still need finalize US cloud migration
- OpenSSL Vulnerability for ATLAS Central Services
  - Scanned all machines
    - older than openssl-1.0.1e-16.el6_5.7 on SL6 vulnerable
    - older than mod-spdy-beta-0.9.4.2-413 is also vulnerable
  - Updating all machines in Quattor or by yum updates on AI nodes
  - Restart apache servers
  - Revoking certificates (but AI nodes not revoked by CA yet)
- Migrating from Savannah to JIRA
  - 26 projects, 7 done, 5 migrating, 9 frozen, 5 for deletion
  - Deadline: May 1
  - New project: database deployment and ops
  - Roles and ACLs a little more complicated in JIRA
  - Don't have CC feature (watchers) at moment - being requested
  - Only active ATLAS members will be able to see tickets as default for all computing projects
- Optimising Software installation setups - development project
  - Mostly for use in HPC environments
  - Some issues with missing setup files in CVMFS for 17.7.3.10+11
  - Setups seem OK in 19.0.1
- Overload of Frontier by MC Overlays
  - Access to multiple connditions instances and pulling in much more data than normal jobs (10-200x)
  - Use either DB release files of Frontier for conditions
    - MC Overlay can't just use default DB release
  - Used to do some of this via HOTDISK but not sure can scale to CVMFS (too many large files?)
  - Maybe use IOVDbSvc - larger time windows
  - Some dependency and assumptions on how MV Overlay tasks are deployed (where is caching done)
  - Bunch-wise lumiosity being accessed but not used
  - No single fix or quick solution - being studied and report back in couple of weeks
- NTUP_COMMON Train Production
  - Derivation Framework is ready to run as pre-DC14 test - finalizing caches and builds now
    - https://twiki.cern.ch/twiki/bin/viewauth/AtlasComputing/CommonD3PD
  - Estimate 300TB data volume, MC TBD
  - AMSG report envisaged O(1TB) per derivation but many are 10x bigger - not clear users will be able to access this efficiently
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