[468] in SIPB-AFS-requests
Ronald-ann backups
daemon@ATHENA.MIT.EDU (Richard Basch)
Wed May 22 14:03:06 1991
Date: Wed, 22 May 91 14:03:47 -0400
To: sipb-afsreq@MIT.EDU
From: Richard Basch <basch@MIT.EDU>
I have the new software suite ready to install...
There are a couple options:
1. Keep the cloning to something that needs to be done on a manual basis.
2. Have Ronald-Ann actually do the cloning
i. Have it done nightly... this way people can recover their previous
day's work.
ii. Have it only done weekly to correspond with how often we actually
spool data to tape.
My personal preference would be to have Ronald-Ann do the cloning on a
nightly basis.
It should log the output of the vos backups so that people can check if
any errors occurred and correct them.
The AFS 3.1 backup software is certainly a lot more robust than the 3.0
Beta software. I have already put in hooks so that it can deal with
other cells in a clean fashion. There are a couple problems with how it
maintains its database, but given that you have different dump schedules
and re-use the tapes associated with a particular dump schedule, there
will be no problems. The problem to which I am referring is that the
backup software deletes all associated tapes from its database having
the same volumeset/dump-schedule marking - it ignores the timestamp on
the tape, since that is not being properly generated (or rather it
doesn't match the database time stamp).
Are there any objections to:
#1: Installing the new backup software?
#2: Installing the automatic cloning software?
Depending on #2, I may have to modify the instructions I wrote up to go
along with the new backup software.
-Richard
