[29873] in Hotline Meeting
AFS resync
daemon@ATHENA.MIT.EDU (dkk@MIT.EDU)
Tue Sep 12 21:05:19 1995
From: dkk@MIT.EDU
Date: Tue, 12 Sep 1995 21:05:07 -0400
To: athena-outage@MIT.EDU
Cc: op@MIT.EDU
I'm planning to re-synchronize the AFS protection database (with the
Moira database) on Saturday morning. The AFS protection service
provides authorization for access to files in AFS. (It keeps track of
the groups you're in.)
The end of this procedure entails restarting the protection servers,
which will result in a period of time (under 5 minutes) during which
attempts to access AFS files will hang. If done at a peak time, it
could cause a meltdown, but as scheduled, it should be only a pause in
service.
I'll be keeping Zephyr instance consult informed during the procedure.
DCNS does not support non-standard AFS groups or other protection
database entities, but most of those that exist will be preserved.
SCHEDULE:
Saturday, 16 September 1995, 6am-9am.
REASON:
To clean up AFS groups and memberships.
SERVICES:
Moira and AFS. This includes any use of moira, blanche
namespace, or any of the *maint programs, and almost all
Athena lockers (and anything which relies on a locker).
IMPACT:
Moira's daily unavailability for backups (usually 5am-~6:15am)
will last longer than usual. AFS will be unavailable for a
few minutes.