[287] in Athena_Backup_System
ABS minutes for 7/11/96
daemon@ATHENA.MIT.EDU (Jonathon Weiss)
Thu Jul 11 16:19:07 1996
Date: Thu, 11 Jul 1996 16:18:52 -0400
To: athena-backup@MIT.EDU
From: Jonathon Weiss <jweiss@MIT.EDU>
Attending: vrt, delgado, miki, wdc, dkk | jweiss
Some discussion about the format of Jeff's review of our code. we
need to make sure we are on the same wavelength as him.
miki: I don't know what I'm supposed to be doing as part of the team,
especially with respect to the code reviews. She sees severall
possible roles. One: the locker orginiation could some work (mostly
documenting what's there, not necessarily changing it). Two: It does
not look like the design document was ever carefully reviewed.
writing an even lower level design doc for the master (to the level of
pseudocode). Look into adding more file system types.
Diane likes the idea of miki taking a large role in the team, looking
at the code and possibly doing cleanup or simplification.
miki: it's not clear how well all of the ABS will map onto things
like NFS, because we don't generally want to restore an entire
partition. We may only want a few files off of a 1G partition that we
backed up. Lots of discussion on how we might want to deal with this.
I think we decided that we liiiike the idea of declaring entities
(equivalent to volumes when using ABS on AFS) to map onto directories
at the top level of the partition, but there are a lot of
implementation details to worry about since we have different usage
patterns between things like root partitions on servers and data
partition on nfs servers. We also don't have a good location system
like the AFS vldb for UFS and NFS.
Those people who are here next week should talk about beta.
