[2588] in SIPB-AFS-requests
Re: New AFS server parts are all here
daemon@ATHENA.MIT.EDU (Jonathon Weiss)
Mon Dec 23 18:51:20 1996
From: Jonathon Weiss <jweiss@MIT.EDU>
To: Greg Hudson <ghudson@MIT.EDU>
Cc: sipb-afsreq@MIT.EDU
In-Reply-To: Your message of "Mon, 23 Dec 1996 15:10:50 EST."
<9612232010.AA25557@small-gods.MIT.EDU>
Date: Mon, 23 Dec 1996 18:47:05 EST
I would suggest the following plan, instead of Greg's:
1) install new machine as sipb-server-1 (same install issues glossed
over as Greg did)
2) Put one or two of the new disks on rc
3) Migrate all data from ra to rc *
4) replace ra and start db processes on it
5) balance the cell as desired
* If a volume has a RO on ra and rc already, the ra RO should be moved
to rb, and an approprioate amount of other data should be moved to rc
instead.
This should shorten the outage to the time it take s to plug a new
disk into rc, and put it in the fstab (we can newfs it ahead of time
on sipb-server-1) and eliminates the need for rebuilding the vldb. It
has the slight disadvantage that we don't keep the 4 new disks on the
same machine, but I don't think that's too big of a deal.
Jonathon
