Red Hat Bugzilla – Bug 240567
if dump-core fails on domain crash domain persists and retries every xm or xenwatch
Last modified: 2010-10-22 11:10:11 EDT
Description of problem:
When a domain crash then if xend cannot dump-core memory of the domain,
the domain is not destroyed, so xend try to dump-core it
every time run "xm list" or xenwatch.
These retry fail if the disk is full, so this issue is occurred.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Configure "dump-enable yes"
2. arrange for /var/lib/xen/dump to have insufficient space.
3. destroy or crash a domain
incomplete core is created on crash and every subsequent xm command.
Core or no core.
This is fixed in an upstream patch (something similar to which is in the F7 xen)
Patch looks sane, although it does not apply to current RHEL5 Xen tree. After a
cursory glance, it should be fairly simple to re-diff it though.
Fujitsu confirmed this issue was gone by 5.1 beta.
I confirmed this issue is fixed in RHEL5.1 beta.
Category set to: Xen::Base OS
Status set to: Next Update
Resolution set to: 'RHEL 5.1'
This event sent from IssueTracker by mmatsuya
Closing per Comment #2 above.