RE: [Xen-ia64-devel] Weekly benchmark results [ww24]

["yo.fujita" <yo.fujita@xxxxxxxxxxxxxxxx>]

Hi Yongkang, Alex,

I tried the latest cset 10419 too.
And the same problem was reproduced.
I think this may be caused by the "credit" scheduler.
Now our developers are researching this problem.

Setting
server      :tiger4
dom0mem     :512M
domUmem     :512M
domU cpus   :2
sched       :credit

Test details
create 3/10 hung with dom0.
destroy 4/7 hung with dom0.

Thanks,
Fujita
> -----Original Message-----
> From: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx
> [mailto:xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of yo.fujita
> Sent: Monday, June 19, 2006 4:21 PM
> To: 'You, Yongkang'; 'Alex Williamson'
> Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> Subject: RE: [Xen-ia64-devel] Weekly benchmark results [ww24]
> 
> Yongkang,
> 
> I have a request.
> I'm using a scheduler "credit" for test.
> If you're using other scheduler, can you try the "credit"?
> 
> Thanks,
> Fujita
> > -----Original Message-----
> > From: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx
> > [mailto:xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of
yo.fujita
> > Sent: Monday, June 19, 2006 4:03 PM
> > To: 'You, Yongkang'; 'Alex Williamson'
> > Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> > Subject: RE: [Xen-ia64-devel] Weekly benchmark results [ww24]
> >
> > Hi Yongkang,
> >
> > Thanks for your information!
> > We also must try the latest changeset.
> > If it happens again, the cause is in our environment.
> > I'll inform you a result. Please wait for a while.
> >
> > Thanks,
> > Fujita
> > > -----Original Message-----
> > > From: You, Yongkang [mailto:yongkang.you@xxxxxxxxx]
> > > Sent: Monday, June 19, 2006 3:37 PM
> > > To: yo.fujita; Alex Williamson
> > > Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> > > Subject: RE: [Xen-ia64-devel] Weekly benchmark results [ww24]
> > >
> > > Hi Fujita,
> > >
> > > I tried the latest Changeset 10419. But I couldn't reproduce this
> problem
> > in
> > > my box. Is it fixed? I create and destroy SMP XenU for more than 100
> > times.
> > > And try to make kernels in xenU 2 times. I didn't meet the xen0 hang.
> > >
> > > My box has 3072M memory. Xen0 has 512M, xenU has 512M. XenU has been
> > assigned
> > > with 2 CPUs.
> > >
> > > Best Regards,
> > > Yongkang (Kangkang) 永康
> > >
> > > >-----Original Message-----
> > > >From: yo.fujita [mailto:yo.fujita@xxxxxxxxxxxxxxxx]
> > > >Sent: 2006年6月19日 10:27
> > > >To: You, Yongkang; 'Alex Williamson'
> > > >Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> > > >Subject: RE: [Xen-ia64-devel] Weekly benchmark results [ww24]
> > > >
> > > >> Hi Fujita,
> > > >>
> > > >> Maybe it is not guest image issue. In our nightly testing for XenU,
> we
> > > >still
> > > >> focus on the UP stability and basic booting/destroying testing.
> > Automatic
> > > >SMP
> > > >> XenU isn't fully added into nightly testing (will add it this
week).
> I
> > > >noticed
> > > >> your report mentioned that the stress testing and keeping
> > > >creating/destroying
> > > >> SMP XenU will catch this issue. I can do some trying to see if I
can
> > > >reproduce.
> > > >Hi Yongkang,
> > > >
> > > >Thanks for your comments.
> > > >As you said, I meant the problem is not in the image itself but the
> > stress
> > > >of booting domU. In other words, I thought larger size image takes
more
> > > >stress on Xen.
> > > >So I guess a customized image for testing (necessity minimum size)
> > doesn't
> > > >cause
> > > >the problem.
> > > >We appreciate it if you can try to reproduce these issues.
> > > >
> > > >Thanks,
> > > >Fujita
> > > >
> > > >
> > > >>
> > > >> Best Regards,
> > > >> Yongkang (Kangkang) 永康
> > > >>
> > > >> >-----Original Message-----
> > > >> >From: yo.fujita [mailto:yo.fujita@xxxxxxxxxxxxxxxx]
> > > >> >Sent: 2006年6月19日 8:53
> > > >> >To: 'Alex Williamson'; You, Yongkang
> > > >> >Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> > > >> >Subject: RE: [Xen-ia64-devel] Weekly benchmark results [ww24]
> > > >> >
> > > >> >>    Thanks Fujita.  Is anyone else seeing these hangs booting
domU?
> > > >> >> The Intel status report on the same changeset has no indication
of
> > > >> >> dom0 hangs on their tests.
> > > >> >Hi Alex, Yongkang,
> > > >> >
> > > >> > In Fujitsu, it's only me who saw this problem because few
> developers
> > > >have
> > > >> >started any tests relating to SMP.
> > > >> >I think the reason why Intel's results was not match to ours is
the
> > > >> >difference of stress on booting SMP domU because the our guest
image
> > is
> > > >just
> > > >> >only a copy of root file system of native Linux, which any
functions
> > was
> > > >not
> > > >> >edited.
> > > >> >
> > > >> >My domU environment.
> > > >> >Disk size :5G
> > > >> >OS        :RHEL4U2
> > > >> >Memory    :512M
> > > >> >CPUs      :2
> > > >> >Yongkang, do you have any comments?
> > > >> >
> > > >> >> Has anything else changed in the test environment?  Thanks,
> > > >> >No. We switched this test to SMP from UP two weeks ago.
> > > >> >
> > > >> >Thanks,
> > > >> >Fujita
> >
> >
> >
> > _______________________________________________
> > Xen-ia64-devel mailing list
> > Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> > http://lists.xensource.com/xen-ia64-devel
> 
> 
> 
> _______________________________________________
> Xen-ia64-devel mailing list
> Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-ia64-devel



_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel