xm migrate --live panics hypervisor
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
xen (Ubuntu) |
Confirmed
|
Undecided
|
Unassigned |
Bug Description
I have to HP Proliant D380 Gen8 Servers with a shared iSCSI storage in a xen cluster configuration. They run 12.04 amd64 and are fully patched and freshly rebooted.
When i call
/usr/sbin/xm migrate --live <some-domain-id> <ip-over-
on one server this "sending" system reboots instantly and the VM only appears in an undefined state on the "receiving" host.
After redirecting the console output of the hypervisor to com1 i was able to retrieve this error message:
(XEN) [2015-11-11 07:50:05] traps.c:3073: GPF (0000): ffff82c48017fa31 -> ffff82c4802070d8
(XEN) [2015-11-11 07:50:15]
(XEN) [2015-11-11 07:50:15] *******
(XEN) [2015-11-11 07:50:15] Panic on CPU 10:
(XEN) [2015-11-11 07:50:15] mm lock held by sh_page_
(XEN) [2015-11-11 07:50:15] *******
(XEN) [2015-11-11 07:50:15]
(XEN) [2015-11-11 07:50:15] Manual reset required ('noreboot' specified)
After replacing "/boot/
uname -srvm
Linux 3.2.0-94-generic #134-Ubuntu SMP Fri Nov 6 18:16:45 UTC 2015 x86_64
Working hypervisor versions:
(XEN) Xen version 4.1.6.1 (Ubuntu 4.1.6.1- 0ubuntu0. 12.04.5) (<email address hidden>) (gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) ) Wed Mar 11 15:00:28 UTC 2015
(XEN) Xen version 4.1.4 (Debian 4.1.4-3+deb7u9) (<email address hidden>) (gcc version 4.7.2 (Debian 4.7.2-5) ) Sat Oct 31 06:17:33 UTC 2015
Non working hypervisor version:
(XEN) Xen version 4.1.6.1 (Ubuntu 4.1.6.1- 0ubuntu0. 12.04.6) (<email address hidden>) (gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) ) Wed Sep 2 18:06:32 UTC 2015