Bug 59551

Summary:	installer crash when 6 packages left
Product:	[Retired] Red Hat Linux	Reporter:	Pekka Pietikäinen <pp>
Component:	anaconda	Assignee:	Michael Fulbright <msf>
Status:	CLOSED RAWHIDE	QA Contact:	Brock Organ <borgan>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	7.3
Target Milestone:	---
Target Release:	---
Hardware:	i386
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2002-02-25 17:21:20 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Pekka Pietikäinen 2002-02-10 08:12:35 UTC

Description of Problem:

When upgrading from 7.2 -> 7.2.90 the upgrade went mostly ok until
there were just 6 packages left to install. Then the installer crashed
and the system went back to text mode where there were no helpful messages
about the reason of the crash.

upgrade.log ended with 

Upgrading wordtrans-web.
Upgrading Xbae-devel.
Upgrading xcdroast.
Upgrading xfig.

(there should have been a few after that like zsh).

After rebooting the system the rpm database was just like before the
update (== all package versions were the 7.2 ones even though the
packages had been upgraded)

This problem _might_ be related to flaky media, as the cd was in my
dvd drive that had some problems reading the disc3 cd-rw reliably
(my plextor read the disk just fine)

Comment 1 Michael Fulbright 2002-02-11 16:06:46 UTC

Did you run 'linux mediacheck' to test the CDs you burned as directed in the
release notes?

Comment 2 Pekka Pietikäinen 2002-02-25 17:21:15 UTC

I did. Can't remember which drive the cd was in during the media check,
it's been two weeks.  In any case the installer shouldn't crash 
like that, 7.2 at least displayed a dialog that offers a retry possibility.

Comment 3 Jeremy Katz 2002-03-14 23:33:47 UTC

depends on the case in which it crashes... sometimes fall over in levels that
don't pass up errors nicely (*cough* the ide/scsi layers *cough* :) or rpm not
passing back a callback.  I've instrumented a new rpm callback that will
hopefully be in beta3 that will catch some of these things more nicely.

The newer rpmdb is in /var/lib/rpm.rebuilddb* btw