Re: [Xen-devel] [xen-unstable test] 13861: regressions - trouble: blocked/broken/fail/pass

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 25.09.12 at 10:44, Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx> wrote:
> xen.org writes ("[xen-unstable test] 13861: regressions - trouble: 
> blocked/broken/fail/pass"):
>>  build-amd64-pvops             2 host-install(2)     broken REGR. vs. 13825
> 
> This is another machine having forgotten that it was supposed to boot
> from the network.  It's itch-mite, the twin of the other test box
> gall-mite which had this problem yesterday.
> 
> That two identical machines have had this happen within days suggests
> that somehow something done to them recently has corrupted their
> CMOS, rather than it being a random event.
> 
> Jan, do you think it at all plausible that the cpuidle-related Xen bug
> is somehow responsible ?  TBH it doesn't seem all that likely because
> I've checked gall-mite and the CMOS is still fine, even though the
> machine has experienced several of these crashes.

No, I don't see how this ought to be connected. Given previous
occasions, is it perhaps that early Xen crashes (maybe in
connection with the way Xen reboots the system afterwards)
cause this independent of the reason for the crash? Or is there
any sort of boot cycle detection active on those systems?

Jan


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