[26180] in Athena Bugs

home help back first fref pref prev next nref lref last post

linux 9.3.11 - 9.3.16: update

daemon@ATHENA.MIT.EDU (Camilla R Fox)
Sat Nov 13 17:30:42 2004

Message-Id: <200411132230.iADMU8Sx021053@red-herring.mit.edu>
To: bugs@mit.edu
Date: Sat, 13 Nov 2004 17:30:08 -0500
From: Camilla R Fox <cfox@mit.edu>
Errors-To: bugs-bounces@mit.edu


System name:		red-herring.mit.edu
Type and version:	i686 9.3.16 (with mkserv)
Display type:		nVidia Corporation NV5 [RIVA TNT2/TNT2 Pro] (rev 15)

Shell:			/bin/athena/bash
Window manager:		ctwm

After a hang (no mouse cursor, pings but nothing else) which I attribute
to the peicewise xscreensaver module, I rebooted red-herring, then tried
to update it from 9.3.11 to 9.3.16.

It spent half an hour appearing to install RPMs normally.  When
it reached 100%, it just sat there.  An hour later, I decided to
investigate, and found that an rpmupdate was running; it had no
children.  I straced it, and found only:
  futex(0x908a004, FUTEX_WAIT, 0, NULL <unfinished ...>

I kill -9'd it, the update failed, and when I reran the update,
rpmupdate quickly removed a list of packages, then exited successfully,
so I assumed all was well and rebooted.  (This all happened on
Thursday.)

Today I noticed that my rpm binary and in fact much of the contents of
the rpm rpm were missing.  (Notably not the contexts of /var/lib/rpm
although the __db files were missing.)

I copied in the missing binaries (only things that I presumed were not
system dependant) and /usr/lib/rpmrc from another system.

I still seem to be sort of screwed, though:
  red-herring.mit.edu# rpm --rebuilddb
  error: unrecognized db option: "db3" ignored.
  red-herring.mit.edu# rpm -qa
  error: cannot open Packages database in /var/lib/rpm
  no packages

Do I have any good options for recovering?  The system is up and
functional, but I'd like to refresh my backups before doing anything
more drastic.

What should I have done when confronted with the hung update?

Jonathon mentioned an update problem on a cluster machine, that sounded
similar; do we have other evidence that this might not be isolated?

-Camilla

home help back first fref pref prev next nref lref last post