Revision as of 10:02, 23 April 2014

RAL Tier1 Operations Report for 23rd April 2014

• We note that Atlas has managed to fill the OPN link inbound to RAL for around 24 hours on 9th April.
• Over weekend (12/13 April) there were problems with our Atlas Frontier Service which went down. This was part of extended Frontier problems that affcted other Tier1 sites too.
• Maintenence on Primary OPN link early evening on Monday 14th April took the link down for a few hours. The failover to the backup link again did not work properly. The effect of this can be seen in the failure of the SUM tests from CERN during this time.

• * GDSS403 (AtlasTape - D0T1) failed on Friday 11th April. At the time we understood there was one file on it waiting to go to tape. Following a drive replacement it was put back into service. However, a further 54 files were then found to be unmigrated to tape and were corrupt. These files have been declared lost to Atlas.

• The load related problems reported for the CMS Castor instance have not been seen this last fortnight. However, work is underway to tackle these problems, in particular servers with faster network connections will be moved into the disk cache in front of CMS_Tape when they become available.
• The Castor Team are now able to reproduce the intermittent failures of Castor access via the SRM that has been reported in recent weeks. Understanding of the problem is significantly adcanced and further investigations are ongoing using the Castor Preprod instance. Ideas for a workaround are being developed.
• As reported before, working with Atlas the file deletion rate was somewhat improved. However, there is still a problem that needs to be understood.
• Problems with the infrastructure used to host many of our non-Castor services have largely been worked around, although not yet fixed. Some additional migrations of VMs has been necessary.
• We have now had repeated instances where the OPN link has not cleanly failed over to the backup link during problems with the primary.
• One of the network uplinks (for the 2012 disk servers) has been running at full capacity. We have a plan to move the switch into the new Tier1 mesh metwork to alleviate this.

• GDSS566 (AtlasDataDisk - D1T0) failed yesterday evening (22nd April). System out of production while investigations continue.

• Appropriate systems patch for the "heartbleed" vulnerbility.
• Environment modification made to enable T2K jobs to run on the ARC CEs.
• Disk usage limits (of 150 GB) added to CEs.
• Alias "cernvmfs.gridpp.rl.ac.uk" updated to point to cvmfs-wlcg.gridpp.rl.ac.uk - the new CVMFS stratum-1 v2.1 on Tuesday (15th April).

Listing by category:

• Databases:
  • Switch LFC/FTS/3D to new Database Infrastructure.
• Castor:
  • Castor 2.1.14 testing is largely complete. (A non-Tier1 production Castor instance has been successfully upgraded.) We are starting to look at possible dates for rolling this out (probably around May).
• Networking:
  • Move switches connecting recent disk servers batches ('11, '12) onto the Tier1 mesh network.
  • Update core Tier1 network and change connection to site and OPN including:
    • Install new Routing layer for Tier1 & change the way the Tier1 connects to the RAL network. (Scheduled for 29th April)
    • These changes will lead to the removal of the UKLight Router.
• Fabric
  • We are phasing out the use of the software server used by the small VOs.
  • Firmware updates on remaining EMC disk arrays (Castor, FTS/LFC)
  • There will be circuit testing of the remaining (i.e. non-UPS) circuits in the machine room during 2014.

Key: Atlas HC = Atlas HammerCloud (Queue ANALY_RAL_SL6, Template 508); CMS HC = CMS HammerCloud