AFS weekly meeting

Monday 16 Mar 2015, 15:00 → 16:00 Europe/Zurich

31/1-012 (CERN)

AFS service/operation meeting

AFS operation meeting 16.03.2015

present: Dan, Kuba, Massimo, Jan

• Dan got called on Saturday for afsacl01 (INC0752722), operator did not know how to reset the VM. Jan to make sure that the procedure is clear on the (un)importance of these VMs.
• Backup/TSM/ABS:
  • ran cleanup 2x, now at "persistent residual level" (ran manually, something still wrong with afscron1 but no longer matters)
  • own Volume accounting: down to 7TB. TSM says 70TB.
  • cleaned up 2 broken CASTOR work dumps form February (both from same day). Just (as_admin) check via "nsls" that these are 0-size and "nsrm", ABS will cope.
• SLS: added new sub-SLSes for user/work/git/svn, started to "push" XML via lemon-agent. Hope to be able to auto-notify the GIT people once their volumes are in trouble.
• new LEMON metric that counts the "rescheduled" lines
• Bug: "No buffer space available" =~ overloaded server+timeout (and bad errno translation)
• RAID firmwares all updated, last was afsdb3
• propose to go for syslog logging, e.g together with CC7. Advantage is longer "memory", and non-local copy.
  • discuss: prefer to keep separate log files (although not in /usr/afs/logs/)
  • affects trace_fs