[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Crashes on image file backed VMs

On Fri, 2015-12-18 at 11:02 +0000, Andre Fucs wrote:
> Â Â Â Â Â Â Â Â Stack trace of thread 2756:
> Â Â Â Â Â Â Â Â #0 Â0x000055b2a1e155d3 blk_handle_requests (qemu-system-i386)
> Â Â Â Â Â Â Â Â #1 Â0x000055b2a1d99fb3 aio_bh_poll (qemu-system-i386)
> Â Â Â Â Â Â Â Â #2 Â0x000055b2a1d99bdc aio_poll (qemu-system-i386)
> Â Â Â Â Â Â Â Â #3 Â0x000055b2a1d99de3 aio_ctx_dispatch (qemu-system-i386)
> Â Â Â Â Â Â Â Â #4 Â0x00007fef67d30dc7 g_main_context_dispatch 
> (libglib-2.0.so.0)
> Â Â Â Â Â Â Â Â #5 Â0x000055b2a1eba35e glib_pollfds_poll (qemu-system-i386)
> Â Â Â Â Â Â Â Â #6 Â0x000055b2a1f2bab2 main_loop (qemu-system-i386)
> Â Â Â Â Â Â Â Â #7 Â0x00007fef6420f610 __libc_start_main (libc.so.6)
> Â Â Â Â Â Â Â Â #8 Â0x000055b2a1d997b9 _start (qemu-system-i386)

This thread looks like the culprit (the rest are all basically idling)

If you run gdb on the binary + core combination you ought to be able to get
more info, like line numbers and perhaps even local variable states which
would maybe point towards the culprit. I don't know if you will also need
to install any debug symbols packages on Arch or not.

One the face of it it doesn't seem to be Xen specific, i.e. it looks like a
QEMU issue (albeit in the version of QEMU shipped along with the Xen

> has anyone seem something like that?

FWIW I don't recall any similar reports and our automated tests include
qcow and they seem pretty happy on the 4.5 branch:
(the failures at the bottom, resolved in flight62168, were a machine
specific timeout, not a crash like this one)

Those tests are passing on that branch after bbbd29a25d09 AKAÂRELEASE-
4.5.2~37, which does suggest you might want to try updating from 4.5.1,
just in case. (These tests are new enough not to have tested 4.5.1)


Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.