Tech Meeting Minutes 20100930

From ScotGrid

Present: Graeme, Sam, Wahid, Mike, Mark, David, Andrew, Stuart

http://indico.cern.ch/conferenceDisplay.py?confId=108379

Backlink: http://www.scotgrid.ac.uk/wiki/index.php/ScotGrid_Technical_Meetings

Table of contents

Security

  • Note that SL4 kernel now available - this will require further patching and reboots of older nodes (lcg-CEs).
  • Glasgow
    • Sam will patch SL5 disk servers and DPM headnode. This will require a very short downtime. Will be scheduled for tomorrow morning.
      • Sam has patched SL5 disk servers and DPM headnode. This required a very short downtime. It was scheduled for this morning.
  • ECDF
    • Patched

Site Issues

  • Durham
    • Need to do a syncat dump of the DPM to clean dark data.
    • Peter reports by email having done this, but it was sent to Alessandra (who's on holiday). Graeme requested that he send it to him or post it onto the savannah ticket: https://savannah.cern.ch/support/?116973.
  • Availability was excellent last week - well done to all!

Site Updates

CREAM:

  • Glasgow: Mark has updated the test CREAM CE at Glasgow. A memory leak was noted - needs to be watched out for. This is experimentally submitting to SGE.
  • ECDF: Andy has a CREAM box installed, currently waiting for ECDF systems team to allow it to submit batch jobs. stdout/err not resolved, but not critical for ATLAS.
  • Durham: needs discussed.

glite-APEL:

  • Low priority action to move from glite-MON to glite-APEL by end of year.
    • Glasgow: On hold.
    • ECDF: On hold.
    • Durham: need discussed.

New Kit:

  • Storage
    • Glasgow
      • Networking: ClusterVision have problems securing the DELL switches. This prevents storage deployment (no bandwidth!). Currently trying to source them from the US.
      • Disks: David and Sam will start to burn them in very soon (local tests). A few single disks rejected.
      • ACTION: David to send burn in scripts to Wahid.
      • WNs: Kit was burned in this week. Good HS06 scores (14.8/core). Now being enabled in the batch system - reservation for atlprd as this has low networking needs.
    • ECDF, Wahid has new disk servers installed. Waiting for ECDF systems team to open network ports.
    • Durham: Have 3 quotes for the tender. Hope to install new kit in the new year.

ARC

  • High priority to install an ARC CE at ECDF to finish CHEP work.
    • ACTION: Stuart and Andy to liaise on this. Stress good publicity to ECDF team to help with any infrastructure issues.

ScotGrid Poster

  • Mike will work on this next week.
    • ACTION: Wahid, Mike to co-ordinate on site specs for the poster. Need inputs from Durham also.

AOB