[1996] in SIPB-AFS-requests
Disk wanted for penguin-lust [pshuang]
daemon@ATHENA.MIT.EDU (pshuang@MIT.EDU)
Thu May 18 03:34:44 1995
From: pshuang@MIT.EDU
Date: Thu, 18 May 95 03:34:19 -0400
To: star-maintainers@MIT.EDU
Cc: pshuang@MIT.EDU
[Comment: please continue to include me in the destination headers if
you reply to this email.]
A late-night discussion with Jonathan Weiss and Matt Powers about rtfm
services migration to penguin-lust suggests a possible resource
problem. After May 26 (thesis due date) but before I start work in
mid-July, I'm commited to make the effort to make everything which
runs on bloom-picayune related to *.answers and the periodic
informational postings archives run on penguin-lust, while leaving
them still running on bloom-picayune; I expect there would be a number
of problems to iron out. After a couple of days of ironing out
problems, *THEN* do an externally visible name cut-over, and a few
days after that, stop what's running on bloom-picayune.
At the current time, there is effectively no free disk space on
penguin-lust to support this activity. Local disk space free is under
500MB, and the rebuilding of the usenet-addresses database comes close
to using up all that space.
Possible solutions:
1) Borrow a disk from I/S temporarily. Cons: too often the loan ends
up being more permanent than either party intended. Even if there
was the best of intentions, this could generate bad feelings.
2) "Borrow" a disk from some other SIPB machine. Any relatively
painless swaps that could net penguin-lust another GB of free
space? (Less additional disk space could be adequate, as the
archives I believe currently take up about 250MB of disk space;
however, I don't believe the swapping strategy is likely to yield a
net gain for penguin-lust in increments less than a GB.)
3) Buy more disk. Someone in the office reported there was about $2K
left in budget which should be spent this fiscal year. Jonathan
guessed off the top of his head that $1,500 would suffice for a 2GB
drive with service warranty --- at this point in time it seems
silly to purchase disk space in smaller increments. (This might
preclude purchasing a larger monitor.)
If we can do (2), that sounds OK. (3) might be the more sensible
thing to do in the long run; if it takes too long to order the disk,
however, it would be a shame to let hardware resource availability
delay the cut-over.
Comments?
--
Ping Huang <pshuang@mit.edu>, who doesn't believe in long elaborate
mail signatures with stuff that most people could care less about....
