Bug 401221

Summary:	stopping ntpd crashes domU
Product:	Red Hat Enterprise Linux 5	Reporter:	David L. Parsley <parsley>
Component:	kernel-xen	Assignee:	Xen Maintainance List <xen-maint>
Status:	CLOSED WORKSFORME	QA Contact:	Martin Jenner <mjenner>
Severity:	high	Docs Contact:
Priority:	low
Version:	5.1	CC:	bdevouge, kbsingh
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2009-02-13 15:49:44 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 David L. Parsley 2007-11-27 16:06:45 UTC

Description of problem:
After running ntpd in a domU for some time (after a week or so), 'service ntpd
stop' crashes the domU kernel.

Version-Release number of selected component (if applicable):
domU:
[root@itfederation ~]# uname -r
2.6.18-53.el5xen
[root@itfederation ~]# rpm -q ntp
ntp-4.2.2p1-7.el5

dom0:
[root@lnxvmhost4 ~]# uname -r
2.6.18-53.el5xen


How reproducible:
Very; it's happened to me 5-6 times on VMs with uptimes > a couple weeks.

Steps to Reproduce:
1. Install a RHEL5 VM w/ ntpd running; don't modify independent_wallclock
2. Wait a few weeks (don't really know how long, but I've not had this problem
on VMs w/ uptimes of only 10-40 minutes)
3. Issue 'service ntp stop'
  
Actual results:
Kernel panic

Expected results:
ntp stops gracefully

Additional info:
ntp is running on dom0.  I've modified our build scripts not to run ntpd on
VMs, since it's unneeded when dom0 runs it.  However, when I went to stop ntp
on a couple of non-critical (fortunately) VM servers today, they both panicked.
 I will try to reproduce this on a test VM and attach the panic output, but I
want to wait a week to be confident it'll happen.

Comment 1 Daniel Berrangé 2008-07-09 12:32:07 UTC

I can't reproduce this problem. If you can still reproduce it on latest RHEL-5.2
kernels, then capture a kernel panic log and attach it to this ticket.

You can easily capture kernel panics by editing /etc/sysconfig/xend and setting
XENCONSOLED_LOG_GUESTS=yes, and then rebooting the host. After this all guest
console output is loged to /var/log/xen/console/guest-NAME.log

Comment 2 Bill Burns 2009-02-13 15:49:44 UTC

Closing this ticket as it was not reproducible and the reported has not responded to the needinfo request for over 6 months. Please reopen is it can be reproduces of you can provide some detail.