● EOS production instances (LHC, PUBLIC, USER)
(Cristi): EOSPUBLIC just got stuck (during the meeting..) after compaction, is now being updated to 4.3.12.
(Luca): EOSHOME instances are on 4.3.12 (except for -4 , waiting for a break in EOSHOME migration/OK from Eddy).
● EOS clients, FUSE(X)
(jan) FUSEX deployment:
- FUSEX still (only) in "qa" for /eos/ams, /eos/lhcb - since Aug 24 (CRM-2790)
- general - what are the criteria to go forward?
- here: push to "production"? add other instances to "qa"?
- AP: Need 4.3.12 client+server for FUSEX, 4.3.11 on FST for FUSEX+FUSE. I.e ATLAS MGM would need to be updated.
- versions:
- 4.3.11 + xrootd-4.8.4 -> "production" (and will go to SLC6/CC7 desktops)
- 4.3.12 -> "qa"
- both might be stuck in KOJI, jan will check.
● Development issues
(Georgios)
- QDB 0.3.3 released: Addition of required commands for running EOS config engine on QDB.
- can run on single node while still in "raft" mode, can then add more nodes (unlike "standalone" mode)
- Q: is backup working - please check with Paul, and try out restore. (Georgios is actually testing this? How can we "test" that the backup is valid? write a new dummy key in REDIS and check for that?)
- eosd private mounts would lose authentication after 20 minutes of inactivity (EOS-2892 - fixed in 4.3.12), caused by two bugs (almost) cancelling each other out.
● AOB
(jan): collecting criteria for restarting the AFS migration, will summarize+send by mail ($HOME probably is last)
There are minutes attached to this event.
Show them.