ABS: some mails form afscron1 - ignored? dangerous? no, look like usual "CASTOR is busy" send2ns() error; we are no longer using a dedicated CASTOR nameserver..
ABS: installed on AFS; worried about ability to disaster recovery situation.. (but currently is on 9 replicas). Should be deployed as RPM; Jan will help with .spec file. ABS needs VLDB to be up for full-machine restores (or to get volume list from AFSCONSOLE DB), could also add fileserver name to CASTOR filename.
ABS: should do a restore exercise at large-but-regular intervals (6months?)
CC7: test VM set up (in testcell), needs newer OpenAFS (CERN 1.6.7 does not cleanly rebuild). Jan to rebase onto 1.6.11
OpenAFS mailing list discussion "OpenAFS still in development?" (admin from TU Berlin looking for fresh Windows builds), confirms overall gloomy feeling
TH/Elena (ticket, triggered by "overload" mails): will talk to her, then organize some "which storage system to use"-tutorial for TH
WEBAFS replacement via EOS: need to talk to AndreasW. Static content can be served from EOS, but .htaccess and cgi-bin need some intermediate web server (which DSS should not run). Ideally could have self-service "EOS-hosted web page" similar to current "AFS-hosted web page"..
needs stable EOS client RPM (still clash with xrootd-4?)
might need a migration campaign for 'default' mapping (http://cern.ch/$USER -> /afs/cern.ch/user/U/$USER/www )
any news on AFS-vs-EOS-FUSE POSIX compatibility test (results)? No.
OpenAFS workshop - send somebody ? IT/DSS travel budget issues, Massimo will cross-check with Alberto [after meeting: no go]
There are minutes attached to this event.
Show them.