AFS weekly meeting

Monday 21 Mar 2016, 16:30 → 16:45 Europe/Zurich

31/1-012 (CERN)

AFS service/operation meeting

AFS operations meeting 2016-03-21

present: Dan, Kuba, Jan

Issues:

• UID changes for ATLAS implicated in getting the ATLAS fileservers stuck last week (Thu/Fri/ Mon?). ATLAS nearly done, CMS to go, (BE delayed at their request)
• ABS errors:  non-backed-up "p.atlas.pro.tzrun2l1" is being copied. If this still cannot back up, will copy content into new volume and rename
• TSM: cleaned out old test volumes and volumes that are (still|again) in CASTOR, left with ~1.5k 'unique' volumes where TSM has the last existing copy. Proposal to delete rejected by Massimo. Proposal to copy into EOS or CASTOR rejected by Jan - if these are really "valuable", would need to properly import into the current ABS naming scheme, and encrypt on tape. Left as-is for now.
• 1.6.17 released, and in our GIT repo. Consider this to be low-impact (no external trust relationship in AD, don't care for random mem leakage over the wire, 1.6.9-cern and above still have a bug in the reschedule path..).
• work.boinc getting stuck now and again. Agreed to move to SSD-backed VM.
• 'uuuafs' account password reset, keytab updated (and docs)
• AFS overload mails - Dan/ Kuba not getting them (or at least not reading, might be filtered..)