AFS operations meeting 2015-04-27
present: Kuba, Dan, Jan (Massimo:excused)
- AFS-on-VM: CEPH now has critical power volumes, so can start to set up AFS servers to use these. First candidates: BOINC, some of the AFS nightly releases?
- could also use upstream AFS there (no thottling) - for small-enough dedicated machines throttling might harm (and BOINC up occasionally)
- ABS: some volumes needed manual unlocking (no "vos move" required), known but low-prio issue
- ATLAS: (email) discussion (triggered by "rescheduling"-nagmail) on p.atlas.pro.tzsoft and "nightly" builds.
- monitor: removed check_ipmi, looking at check_initrd (needs LEMON help)
- enabled "rescheduling"-nagmail for user.X and work.X (works, we've got tickets)
- arc: seems to be client firewall bug (-> RedHat) triggered by half-shutdown socket, will try to rebuild wokaround version from SRPM. Maintainership unclear.
- OpenAFS "best practices workshop" in August (Pittsburgh) - do we go, what would we present?
- afs_admin mildly broken ("-s" option was used for some
obscure volume colocation subcommand, which ATLAS still uses. But not as intended). Fixed.
There are minutes attached to this event.
Show them.