[1176] in Release_7.7_team
Re: mkserv ops and attaching system packs.
daemon@ATHENA.MIT.EDU (Matt Braun)
Fri Dec 19 19:08:02 1997
Date: Fri, 19 Dec 1997 19:07:59 -0500 (EST)
From: Matt Braun <mhbraun@MIT.EDU>
To: Bill Cattey <wdc@MIT.EDU>
Cc: release-team@MIT.EDU, kelley@MIT.EDU
In-Reply-To: "[1175] in Release_7.7_team"
Bill,
The basic idea here is that servers should not have any unnecesary
dependancies, especially AFS dependancies.
If system packs are attached then path searches will go into AFS and it is
more likely that the machine will hang for the 60 second timeout if an AFS
server goes down.
To change this behavior, all you need to do is set RVDCLIENT True in rc.conf
(and put something in /var/server/.private that keeps that change from being
lost).
>Or is this REALLY the right way to do this, and we just need to be a
>little more
>careful in documenting how this works for someone using the Athena
>install to set up a server?
There are several things that mkserv ops does that someone like Peter might
not like. Keep in mind that we wrote it to fit our needs over the years.
There are probably several things that 'mkserv ops' does that a customer like
Peter might not want. For example, we set MRUPDATE true, which essentially
gives anyone with root on moira root access on the server. It also (as you
know) configures the machine to centrally syslog to syslogger.mit.edu, which
might not be what a customer (or ops) wants.
I think the summary here is that using Athena machines as servers is not much
different from using vanilla OS's, in both cases if you don't understand the
assumptions that were made in configuration, then you will likely shoot
yourself in the foot. If you and Peter want to use 'mkserv ops' for servers
that are maintained by someone other than ASO, then maybe you should sit down
with someone in ASO and make sure you know what 'mkserv ops' does.
Matt
