[2246] in SIPB_Linux_Development
Re: Linux Athena questions
daemon@ATHENA.MIT.EDU (Salvatore Valente)
Sun Oct 18 01:42:52 1998
Date: Sun, 18 Oct 1998 01:42:35 -0400 (EDT)
To: tb@MIT.EDU
Cc: linux-dev@MIT.EDU
In-Reply-To: <199810161831.OAA06978@pusey.mit.edu> (tb@MIT.EDU)
From: Salvatore Valente <svalente@MIT.EDU>
[Cc'd to linux-dev because it's fun to reminisce sometimes, and since
there are some good questions here.]
Hi. You wrote:
What were the original reasons for basing the SIPB Linux Athena on the
Redhat distribution?
SIPB Linux Athena originally ran on SLS. Eventually, we (linux-dev)
switched from SLS to Slackware (as did the rest of the world) because
Slackware was just plain better. We switched from Slackware to RedHat
because RedHat had a more intelligent and configurable installation
system, and was quicker to respond to security holes and bugs.
(Slackware took forever to respond to security holes, and there were
lots of them back in the old days.)
What are the technical considerations for continuing to use Redhat
instead of another distribution (especially Slackware, Debian, SuSE,
or any important ones I've left out)? (Ignore the cost of conversion
here; pretend you were starting all over from scratch today.)
I can't ignore the cost of conversion. It's huge.
Back when we (linux-dev) decided to ditch Slackware, RedHat was the
only other popular distribution, so it was a no-brainer. There are
other popular distributions now, but I haven't bothered to look at
them. They may be better, but I have no way of knowing that, and no
real reason to find out. RedHat is sufficient. (Hey, that would be a
good motto! "RedHat: It's sufficient.")
What parts of Athena do not currently work well on Linux boxes?
(Please be brutally honest, it won't be held against you.)
That is, of course, a matter of opinion. My opinion: After all these
years, Linux is still too much of an always-on-the-bleeding-edge
system to work well as a build-once-and-run-for-ten-years system like
Athena. The fact that there's been a new major version of the shared
C library about once a year every year that it's existed makes me
want to either laugh or cry. Locker maintainers shouldn't have to
deal with that. (Neither should users, for that matter.) Linux needs
to stabilize a lot, and it needs to do so about three years ago.
And then there's the AFS cache corruption bug.
Who else should I ask these questions?
linux-dev@mit.edu. Specifically, warlord, amu, and sit have been
active in RedHat Linux-Athena development.
What other questions should I ask to better understand the subject?
I think you've asked the right questions. I expect that some people
will have quite a bit to add to my answers.
Have a nice day,
Sal.
