[1707] in SIPB-AFS-requests
Re: rosebud2 fell over
daemon@ATHENA.MIT.EDU (Derek Atkins)
Tue Jan 17 12:03:08 1995
To: Greg Hudson <ghudson@MIT.EDU>
Cc: sipb-afsreq@MIT.EDU
In-Reply-To: Your message of "Mon, 16 Jan 1995 20:27:11 EST."
<199501170127.UAA15035@lola-granola.MIT.EDU>
Date: Tue, 17 Jan 1995 12:02:40 EST
From: Derek Atkins <warlord@MIT.EDU>
> Maybe we should have a third AFS server running just db server
> processes on, say, bloom-picayune after we get rid of rosebud, in
> order to avoid having clients hang during outages? It wouldn't be in
> the client CellServDB, just the server ones. This plan would raise
> some administration issues and complicate the bloom-picayune setup,
> but I think it's better than the alternative. (The AFS 3.3a servers
> might also fix the bug, so this might be unnecessary, but we have no
> particular reason to believe that.)
While this sounds like a good idea, you have to remember that anyone
with root on an AFS server gets root in the Cell. Therefore, putting
a server on Picayune would be, IMHO, a _bad idea_, unless you want to
remove all other services from Picayune.
On the other hand, having an RT as a backup server, which no one logs
in on, would probably be reasonable... And I don't even have to patch
the AFS code, since it wouldn't be serving files!
-derek
