Re: [Xen-devel] [xen-unstable test] 5506: regressions - FAIL

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

George Dunlap writes ("Re: [Xen-devel] [xen-unstable test] 5506: regressions - 
FAIL"):
> On Tue, Feb 1, 2011 at 12:59 AM, xen.org <ian.jackson@xxxxxxxxxxxxx> wrote:
> >  test-amd64-xcpkern-i386-xl-credit2 11 guest-localmigrate   fail REGR. vs. 
> > 5503
> 
> This one I can't see anything wrong.  The host boots fine, but then
> the ssh to start the localmigrate command just fails.  Nothing on the
> serial console to indicate what might be wrong.

The system seems to have completely hung at the start of the
migration.  The first attempt to capture logs after the test fail is
unsuccessful; it only works after the test script power cycles the
host.

> >  test-i386-i386-xl            14 guest-localmigrate/x10     fail REGR. vs. 
> > 5503
> 
> This one looks a bit like the above, but the ssh failure happens on
> the 3rd iteration; and there are all kidns of scary messages from the
> blk subsystem:

Yikes.

> Jan 31 18:55:44 [  549.991570] ata1.00: exception Emask 0x0 SAct
> 0xffff SErr 0x0 action 0x6 froz

This is at exactly the same point as the above migration failure.  I'm
not sure I would expect the migration to be stopped by failed accesses
to the disk: the sending and receiving processes have both started up
successfully and everything they need should be in the buffer cache.

Is it at all possible that some kind of bug somewhere (in Xen or dom0)
would produce these symptoms ?

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel