Bug 429014

Summary: [RHEL5.2 kdump] ia64 systems unable to boot into kdump kernel
Product: Red Hat Enterprise Linux 5 Reporter: Jarod Wilson <jarod>
Component: kexec-toolsAssignee: Neil Horman <nhorman>
Status: CLOSED ERRATA QA Contact:
Severity: medium Docs Contact:
Priority: medium    
Version: 5.2CC: dlehman, lwang, qcai
Target Milestone: rcKeywords: Regression
Target Release: ---   
Hardware: ia64   
OS: Linux   
Whiteboard:
Fixed In Version: RHBA-2008-0313 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2008-05-21 15:23:16 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 Jarod Wilson 2008-01-16 19:48:35 UTC
Description of problem:
The update from kexec-tools 1.101 + our myriad of patches to tracking the
upstream-ier kexec-tools testing tree has left ia64 systems unable to use kdump.
The crash kernel loads fine, but systems simply hang when they should be booting
into the kdump kernel.

Version-Release number of selected component (if applicable):
kexec-tools-1.102pre-8.el5

How reproducible:
Set up kdump on ia64, echo c > /proc/sysrq-trigger

Additional info:
The latest ia64 5.2 beta kernels still kdump just fine with the 5.1 kexec-tools.

Comment 1 Neil Horman 2008-01-16 21:31:52 UTC
I'm currently trying an upstream tree to see if latest kerenls can boot ia64 via
kdump



Comment 2 Neil Horman 2008-01-17 20:14:21 UTC
Upstream kernels seem to be doa with our kexec-tools as well.  Tring the latest
kexec-tools git tree


Comment 3 Neil Horman 2008-01-17 21:23:45 UTC
note to self: replacing common kexec-elf* files in kexec subdir from 1.101 to
1.02pre fixes this.  I'll need to investigate further.

Comment 4 Neil Horman 2008-01-18 19:31:24 UTC
I've noted two things today.  I can build locally and produce a working kexec
binary that successfully boots a kdump kernel if:
1) I change build_mem_notes to return -1 if any elf headers we read in are
non-null terminated
2) If I don't install the latest elfutils-devel-static on the local system

I have no idea why this works yet.


Comment 5 Neil Horman 2008-01-18 20:15:11 UTC
correction, apparently the actual problem is with the %configure script for the
srpm .  If I forgo the %configure sciprt, and just use ./configure instead, the
rpm built locally works fine.

Comment 6 Neil Horman 2008-01-18 21:10:00 UTC
Narrowing it down further, it appears that the default values for $RPM_OPT_FLAGS
seem to be having an adverse affect on the build for ia64.

Comment 9 Neil Horman 2008-01-21 19:21:12 UTC
Fixed in kexe-tools-1.102pre-8.1, on my private pre53 branch.  I'll merge when
we open for 5.3 checkins.

Comment 11 Neil Horman 2008-01-21 19:46:27 UTC
good point, I misread what version we were shipping in the initial report.
setting exception flag.


Comment 13 Jay Turner 2008-01-29 13:14:49 UTC
QE ack for RHEL5.2.  This is a regression relative to 5.1 and would seriously
impact customer's ability to debug problems in addition to their ability to
provide us with debugging information.

Comment 14 RHEL Program Management 2008-01-29 13:17:27 UTC
This bugzilla has Keywords: Regression.  

Since no regressions are allowed between releases, 
it is also being marked as a blocker for this release.  

Please resolve ASAP.

Comment 19 errata-xmlrpc 2008-05-21 15:23:16 UTC
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on the solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHBA-2008-0313.html