Skip to content

OSG Technology Area Meeting, 26 June 2017

Coordinates: Conference: 719-284-5267, PIN: 57363; https://www.uberconference.com/osgblin

Attending: BrianL, Carl, Derek, Marian, Mat, Suchandra, TimC, Vaibhav

Announcements

  • TimT, Edgar out until next week
  • Derek will be 50% OSG starting July 1, down from 75%

Triage Duty

  • This week: BrianL
  • Next week: Carl
  • 8 (+0) open tickets

JIRA

# of tickets Δ State
149 −7 Open
18 +5 In Progress
6 +3 Ready for Testing
1 +0 Ready for Release

Release Schedule

Version Development Freeze Package Freeze Release Notes
3.4.1 / 3.3.26 2017-06-26 2017-07-03 2017-07-11 Independence Day
3.4.2 / 3.3.27 2017-07-24 2017-07-31 2017-08-08

Notes: Additional “urgent” releases may be scheduled for the 4th Tuesday of each month. The Testing date is when acceptance testing will be scheduled for releasable packages; if a package is added after this date, it may not be possible to schedule adequate testing time, thereby forcing it into the next release.

OSG Software Team

  • Dev freeze today:

    Owner # open tickets
    BrianL 3
    Carl 2
    Mat 2
  • Suchandra: How's the LCMAPS VOMS transition going?

  • HTCondor-CE whole node accounting + memory request issues
  • VMU tests: Erin investigating network issues with new RHEL VMs using the dev subscription. Dakota was brought up to speed on image generation/troubleshooting.
  • ITB progress: New pool configs sprayed out for separated CM and an additional CE for testing pre-release
  • Internal doc migration started and lives here (formerly https://github.com/brianhlin/technology/tree/internal_migration)
  • xrootd-cmstfc (contrib) fails to build for EL7, need help with CMake to place libs in /usr/lib64 rather than /usr/lib/

Discussions

  • RSV JIRA ticket incoming to fix querying condor_q output
  • Incoming Gratia probe changes for whole node accounting issues by Derek, Carl to review
  • Hosted CE LCMAPS VOMS transition complete

Support Update

  • FIT (BrianL) - Jobs held, passed on troubleshooting doc
  • TAMU (BrianL) - SAM tests are better with updates to their scheduler config. Jobs submitted to their backend condor are being held with "job not found"
  • UNL (Derek) - Reprocessed Accounting records for whole node jobs. Pull request for HTCondor-CE (https://github.com/opensciencegrid/htcondor-ce/pull/151) and Gratia-Probes (https://github.com/opensciencegrid/gratia-probe/pull/18)
  • Syracuse & UCSD (Derek) - GPU nodes are starting up. Small amount of support last week, but this week I expect production use to start, so possibly some user support.

OSG Release Team

  • Tim Theisen is handling the July 11th release
3.3.26 Δ Both Δ 3.4.1 Δ Total Δ Status
0 1 −1 0 −5 1 −17 Open
0 −1 6 +4 0 −1 6 +2 In Progress
1 +1 6 +3 3 +3 10 +7 Ready for Testing
0 +0 0 +0 0 +0 0 +0 Ready for Release
0 +0 0 +0 0 +0 0 +0 Closed
1 −1 13 −4 3 −3 17 −8 Total

Ready for Testing

  • Condor 8.6.4 and 8.7.2 in 3.4 and upcoming, respectively. Including af ix for HTCondor-CE-Bosco without certs
  • blahp fix for multicore requests for SLURM batch systems
  • New package, gridftp-dsi-posix, to replace xrootd-dsi
  • Fix for GridFTP startup to use the correct plugin configuration
  • Fix for CVMFS client failing to mount when very large groups exist
  • Added ability to include arbitrary ClassAd attributes in Gratia records
  • Added osg-configure-misc dependency to osg-gridftp
  • Fix for HDFS NameNode infinite loop

Discussions

None this week

OSG Investigations Team

Lots of vacation from Investigations team this week. Not much to update.

Last Week

  • Reprocess underreported sites and upload new APEL report to WLCG
  • Debug Nova CVMFS repo, will need to redo repo with chunking.
  • Packaging of CVMFS-Sync and configurations. https://github.com/bbockelm/cvmfs-sync/pull/1
  • Investigate backups of GRACC peripheral services

This Week

  • Will need to redo nova repo with chunking.
  • Docker'ification of GRACC Agents
  • Start backups of grafana configurations (dashboards and datasources)

Ongoing