[1258] in Moira
boot dcm lossage (2 bugs)
daemon@ATHENA.MIT.EDU (Jonathon Weiss)
Fri Mar 19 17:57:29 1999
Date: Fri, 19 Mar 1999 17:57:24 -0500 (EST)
To: bug-moira@MIT.EDU, moira-admin@MIT.EDU
From: Jonathon Weiss <jweiss@MIT.EDU>
Mike started asking me some questions about the meadow vs. sipb
queues. for those unfamilar with it the current *very unsupported*
situation is that there is a print queue named meadow, spooled by
charon to the printer meadow-p and an SAP print queu named sipb
spooled by arbor-eater also to the printer meadow-p. Concerns about
lossage whe both servers try to print at once, I've discovered several
bugs in the boot dcm.
First, in this situation we find that there are two different files
named /var/boot/hp/meadow-p in the boot.out tarball:
moira3.mit.edu# tar tvf /moira/dcm/boot.out | grep meadow
-rwxr-xr-x 0/0 188 Jun 2 05:24 1993 /var/boot/hp/meadow-p
-rwxr-xr-x 0/0 142 Mar 5 06:24 1999 /var/boot/hp/meadow-p
the right thing here is probably to combine the two entries somehow
into one file in /var/boot/hp, but I'm not sure how hard that will be.
certainly we can merge the allow lists, but I'm not sure what to do
if, say, the trap-dest differs between the two.
Second, we were wondering why both ae and charon were still allowed to
print to meadow-p, since we'd expect that only the one in the second
of the 2 files (ae) in the tarball would work. However, they seem to
both work, because the /var/boot/hp/meadow-p file on tkl corresponds
to neither of the files in the tarball, but is the file from before we
incorporated the printer stuff into moira:
tae-kwon-leap.mit.edu# ls -l meadow-p
-rw-r--r-- 1 root other 5854 Apr 14 1998 meadow-p
We're going to want to be careful deploying the fix to this bug, since
it will change the boot info for a lot of printers.
--
Jonathon
