[708] in SIPB-AFS-requests
AFS disk-space needs for next 3 years?
daemon@ATHENA.MIT.EDU (mhpower@Athena.MIT.EDU)
Thu Jun 25 21:05:13 1992
From: mhpower@Athena.MIT.EDU
To: sipb-afsreq@Athena.MIT.EDU
Cc: sipb@Athena.MIT.EDU, sipb-prospectives@Athena.MIT.EDU
Date: Thu, 25 Jun 92 21:02:20 EDT
Rosebud became an AFS server on September 11, 1991. Before this, we
were using most of the disk space on ronald-ann, maybe about 1.3 Gb.
Currently, we're using about 2.6 Gb out of the 3.0 Gb total (across
the two servers).
This means our disk usage has gone up 1.3 Gb in 9.5 months. Maybe 30%
of this can be explained by the migration of the sipb locker. Another
25% or so is explained by recent large projects such as gnu and tex,
for which much of the disk-space needs may be temporary.
The current "SIPB Hardware Requirements and Machine Proposal" doesn't
request any more disk space over the next three years (7/92 - 7/95).
Actually, it specifies a net reduction of 0.4 Gb: we've asked for a
new AFS server with 1 Gb, to arrive when ronald-ann is decommissioned
in July 1993.
Some ideas we might consider:
-- arrange some yearly allotment of new disk space to SIPB, to be
funded by jdb's office. That is, SIPB should normally expect
to add some amount (probably between 0.5 and 1.5 Gb) of disk
space to the sipb AFS cell each year.
-- if necessary, indicate some minimum fraction of new disk space
that will directly support new software/services for use
throughout the community. This might be something like 85-95%.
-- if useful, indicate some fraction of the new disk space that
we'd be expected to assign to non-members' projects, assuming
there were requests from non-members. SIPB would then be
handling volume creation, cloning, tape backup, etc. for
projects that weren't quite vital enough to be directly
supported by Athena Systems Support.
-- if we can't arrange for much additional disk space, we should
see if it's possible to move some important volumes to the
athena cell, so that we'd retain some reasonable amount of
free space for new/experimental projects. This would also
better serve the users (24 hour operations coverage for the
athena cell, etc.).
Matt
