What kind of hardware is rtfm.mit.edu, and what is your estimate on
how much more gear would be needed to bring things back up to a stable
state?
Rtfm.mit.edu is a DECstation 500/25 with 24Mb of RAM and one SCSI
controller.
The DECstation 500/25 is simply not intended to be a "server model" --
among other things, it can't really hold enough RAM to be able to
handle without swap-thrashing *lots* of daemon processes all trying to
do stuff at the same time.
A more pumped DECstation (e.g., a 500/240), or an DEC Alpha or a
SPARCstation 10, with a good amount of RAM (like, at least 64Mb, and
that's a minimum), would probably be sufficient for the current load
on the machine. However, as someone pointed out on this list just the
other day, the usage of services like this is constantly increasing,
so who knows how long that would last. One of the problems we're
running into now is that the machine we have is just about maxed out;
e.g., I don't think it can take much more RAM.
The SIPB has been talking to MIT Information Systems (IS), our parent
organization, about getting better hardware for our servers.
Unfortunately, the SIPB is a volunteer student organization, which
means that (a) we have high turnaround of officers so it's hard to
keep a solid relationship with the people in IS who matter; (b)
students often have more pressing time commitments which prevent them
from doing SIPB-related work such as writing up hardware proposals,
campaigning to get the hardware, etc.; and (c) as a student
organization, we're sometimes not taken as seriously as we'd like
(however, that isn't *much* if a problem, because of the number of
MIT/SIPB alumni who work for IS :-).
Jonathan Kamens | OpenVision Technologies, Inc. | jik@security.ov.com
[
Usenet Hypertext FAQ Archive |
Search Mail Archive |
Authors |
Usenet
]
[
1993 |
1994 |
1995 |
1996 |
1997
]
© Copyright The Landfield Group, 1997
All rights reserved