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

Re: [Xen-devel] [xen-unstable bisection] complete test-amd64-i386-rhel6hvm-intel



On Thu, 2011-09-01 at 18:22 +0100, Laszlo Ersek wrote:
> On 09/01/11 18:26, Ian Jackson wrote:
> 
> >> job test-amd64-i386-rhel6hvm-intel
> 
> >>    changeset:   23802:bb9b81008733
> >>    user:        Laszlo Ersek<lersek@xxxxxxxxxx>
> >>    date:        Wed Aug 31 15:16:14 2011 +0100
> >>
> >>        x86: Increase the default NR_CPUS to 256
> >>
> >>        Changeset 21012:ef845a385014 bumped the default to 128 about one 
> >> and a
> >>        half years ago. Increase it now to 256, as systems with eg. 160
> >>        logical CPUs are becoming (have become) common.
> >>
> >>        Signed-off-by: Laszlo Ersek<lersek@xxxxxxxxxx>
> >
> > My bisector is pretty reliable nowadays.  Looking at the revision
> > graph it tested before/after/before/after/before/after, ie three times
> > each on the same host.
> >
> > This change looks innocuous enough TBH.  Is there any way this change
> > could have broken a PV-on-HVM guest ?  Note that RHEL6, which is what
> > this is testing, seems to generally be full of bugs.
> >
> > If the problem is indeed a bug in the current RHEL6 then I will add
> > this test to the "do not care" list.
> 
> In what way was the guest broken? How many physical cores/threads was 
> the hypervisor running on?

This is just confusion over the way the failure is reported. The
bisector was running the test-amd64-i386-rhel6hvm-intel job but it was
actually failing at the build/install Xen stage and not getting anywhere
near actually testing rhel6hvm. This confused me (and apparently IanJ)
too. For future reference the thing to look at is the report's header
which in this case said:
        job test-amd64-i386-rhel6hvm-intel
        test xen-install
i.e. the xen-install stage failed while running the
test-amd64-i386-rhel6hvm-intel sequence.

The selection of the test-amd64-i386-rhel6hvm-intel sequence for
bisecting is apparently just an arbitrary choice out of all the
sequences which suffered this failure.

The actual fix for this issue was identified and posted in the "8803:
regressions - FAIL" thread.

Ian.


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


 


Rackspace

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