Description of problem: While running the Xen-domu-stress-save-restore.c on a 64-bit F-11 domU in the Fedora Virt Test Day (step 6) here: https://fedoraproject.org/wiki/QA:Testcase_Virtualization_XenDomU_Save_and_restore, I encountered a failure after a few save/restore iterations. It's not entirely clear to me at this point whether it is a dom0 bug or a domU bug, but since this works on the same dom0 with a RHEL-5 domU, I'll start with domU. The only clue I have currently is in the /var/log/xen/xend.log from the dom0; when it failed, it printed this to the log: [2009-05-07 07:27:53 xend 4071] DEBUG (XendCheckpoint:217) [xc_restore]: /usr/lib64/xen/bin/xc_restore 25 45 1 2 0 0 0 [2009-05-07 07:27:53 xend 4071] INFO (XendCheckpoint:353) xc_domain_restore start: p2m_size = 7ff00 [2009-05-07 07:27:53 xend 4071] INFO (XendCheckpoint:353) Reloading memory pages: 0% [2009-05-07 07:28:29 xend 4071] INFO (XendCheckpoint:353) Received all pages (0 races) [2009-05-07 07:28:29 xend 4071] INFO (XendCheckpoint:353) ^H^H^H^H100% [2009-05-07 07:28:29 xend 4071] INFO (XendCheckpoint:353) Memory reloaded (267221 pages) [2009-05-07 07:28:29 xend 4071] INFO (XendCheckpoint:353) ERROR Internal error: User PT base is bad. pfn=490978 nr=524032 type=00000000 40000000 [2009-05-07 07:28:30 xend 4071] INFO (XendCheckpoint:353) Restore exit with rc=1 [2009-05-07 07:28:30 xend.XendDomainInfo 4071] DEBUG (XendDomainInfo:1905) XendDomainInfo.destroy: domid=45 [2009-05-07 07:28:30 xend.XendDomainInfo 4071] ERROR (XendDomainInfo:1915) XendDomainInfo.destroy: xc.domain_destroy failed.
OK. It looks like this is actually a hypervisor bug, and should be already fixed in the upstream Xen HV. This is also a problem in the RHEL-5 HV, which is being tracked in 494876. I'll close this as a dup of that. Chris Lalancette *** This bug has been marked as a duplicate of bug 494876 ***