Notes:

Infrastructure:

• FNAL has been having issues.
  • Some failures, some low CPU efficiency jobs.
• Vanderbilt had an extended downtime.
• MIT has fallen out of the ITB redirector. Not sure why they haven't been able to keep up?
  • Needs to report DNs; uses grid-mapfile.
• UFL upgraded to Xrootd 3.1
  • 3.0.x servers still left?
• Nebraska has been fine, modulo yesterday's downtime
• UW still needs to enable GUMS integration.
• Higher error rates: seems to be due to glidein misconfiguration?
  • Some low efficiency today, needs to be investigated.
• Disk fill-up on ML repo. Lost 3 days of data.
  • Need to tune to reduce the data volume - larger hard disk has bought us some time.
  • Maybe new nagios test is needed for ML functionality?

Software:

• Need to deploy version of Xrootd where grid map file reports DN.
• xrootd 3.1 was released to osg.
• Monitoring:
  • Added reporting of file size of file close.
  • Time out user sessions.
  • Working on build system.
  • Will transfer the accounting probe to UCSD
  • Still need to get redirection reporting.
  • Error reporting? It's only post-3.1. Also waiting for upgrade.
  • Need to do UDP packet forwarder.
  • Over xmas, plan on doing some realtime viz.
• Caching proxy: Matevz needs to clear off plate a bit for this.
• CVMFS/Parrot: Dan is starting to think about this.
  • Converted the T2 to CVMFS for OSG_APP; got much experience with CVMFS.
  • Think we'll use the FUSE entry points, but not re-implement the protocol. Dan is still trying to get a good feeling based on code.
• MT and BB discussed the possibility of adding the job runtime data to the opaque info that the redirector gets would be much easier to correlate data access and batch jobs.

AOB:

• Last meeting for the year!
• Still need to go through worklist from kickoff meeting.
• Ruth (OSG) asked if AAA kept a blog. We don't yet, but this was the second question about an improved web presence.
  • Something to think about. Maybe delay until next year?