[373] in Athena_Backup_System
Athena Backup System Decision Meeting
daemon@ATHENA.MIT.EDU (Bill Cattey)
Mon Mar 3 14:35:21 1997
Date: Mon, 3 Mar 1997 14:35:14 -0500 (EST)
From: Bill Cattey <wdc@MIT.EDU>
To: athena-backup@MIT.EDU
On Friday 21 February 1997 we held a meeting to decide the fate of the
Athena Backup System Project.
Present were:
Bill Cattey, Ted McCabe, Miki Lusztig, Jonathon Weiss, Mike Barker, Bob
Ferrara, Roger Roach, and Jeff Schiller.
We reviewed the differences between Transarc AFS 3.4a backup, and ABS.
We asked:
What is the relative importance of:
Incremental Backup?
Ability to add new filesystems in the future?
How well tested? (confidence level)
Suport resources spent by us?
Development resources (including time elapsed) to be spent by us?
Richness of information kept by the system?
----
Consensus around the table was:
The richness offered by ABS was not significantly enough more useful
than that maintained by the Transarc backup.
The most important thing was to be able to keep things going as well as
they are now or better.
The most important new functionality would be to cut the backup cycle to
reduce the amount of data lost by customers between dumps. For this
reason, the Incremental Backup item above was the most important.
There is the risk that a few years from now MIT has a different
filesystem than AFS, and that the filesystem that comes with it is not
good enough. This risk is answered if ABS to be shut down gracefully so
that the project could be picked back up later if necessary. We agree
that the work is currently well enough documented so that essentially no
work is required to mothball it.
The current operations folks, Ted and Jonathon specifically, feel that
their requirements will be met by the Transarc AFS 3.4a backup system.
Therefore, we have decided to discontinue the ABS project and proceed to
deploy Transarc AFS 3.4a backup instead.
A project plan to do that is required.
There *IS* the deadline that that a new backup system must be online
before we go live with AFS 3.4a on file servers.
Our organization would benefit greatly by a post-mortem on the ABS
project.
Action Items:
Produce project plan for deploying Transarc AFS 3.4a backup system.
Perform post mortem on ABS project.
The SPARC 5, Turbodog, is to be reclaimed by ops.
The ABS Cell will be reallocated for other work.
OPS-5 has been reallocated for Project Database development work.
The SPARC 20 Metamucil will be reclaimed by OPS for other testing
work.
