Re: [Xen-devel] [xen-unstable test] 108068: regressions - FAIL

[Paul Durrant <Paul.Durrant@xxxxxxxxxx>]

> -----Original Message-----
> From: Ian Jackson [mailto:ian.jackson@xxxxxxxxxxxxx]
> Sent: 02 May 2017 15:04
> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
> Cc: 'Jan Beulich' <JBeulich@xxxxxxxx>; Wei Liu <wei.liu2@xxxxxxxxxx>; xen-
> devel <xen-devel@xxxxxxxxxxxxxxxxxxxx>
> Subject: RE: [Xen-devel] [xen-unstable test] 108068: regressions - FAIL
> 
> Paul Durrant writes ("RE: [Xen-devel] [xen-unstable test] 108068: regressions
> - FAIL"):
> > It's very hard to say. There's nothing particularly conclusive in the logs.
> QEMU has definitely requested a reset as can be seen in
> >
> > http://logs.test-lab.xenproject.org/osstest/logs/108068/test-amd64-i386-
> xl-qemut-winxpsp3-vcpus1/italia0---var-log-xen-qemu-dm-
> win.guest.osstest--incoming.log
> >
> > but really no clue why. The most obvious reason would be a BSOD in the
> OS, but without catching it in the act there's not much we can tell.
> 
> Does a Windows BSOD not crash or hang the VM ?  Rebooting as a result
> is pretty unhelpful in our context (and many others I guess).

Unless you modify the config of the OS (as XenRT does) then the default action 
is to reboot. I assume this is to attempt to maintain availability on the 
normal case.

  Paul

> 
> > IMO testing OS that are so massively out of date is probably not
> worthwhile. If there is any real problem then I'd expect it to manifest across
> multiple versions of Windows and so other tests should catch such an issue.
> 
> I don't have an opinion about this.  I can drop this test.
> 
> Does anyone object ?
> 
> Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel