Ah, thanks for that Alastair. Taking a closer look, it turns out that
the misbehaving WNs missed out on the latest round of glite updates
(which fixed this problem) for some reason. I'll get this sorted out
now.
Cheers,
Rob
> -----Original Message-----
> From: Testbed Support for GridPP member institutes
> [mailto:[log in to unmask]] On Behalf Of Alastair Dewhurst
> Sent: Tuesday, July 28, 2009 2:26 PM
> To: [log in to unmask]
> Subject: Re: UKI HammerCloud Test 2: (Tuesday 28th July):
> This Time It's Personal
>
> Hi
>
> The problem at RALPP is caused by the latest version of gLite
> WN code. This is a problem Graeme spotted around 1 week ago
> and emailed out a link to the work around. I have included
> that email with a link to the work around at the bottom of this email.
>
> Other than this small problem the Hammer Cloud test has
> started extremely well.
>
> Alastair
>
>
> On 28 Jul 2009, at 12:21, Harper, RM (Rob) wrote:
>
> > We've had a bunch of jobs fail at RALPP which, according to Panda,
> > have all thrown an error along the lines of:
> > Error details: pilot: Expected output file
> > user09.JohannesElmsheuser.ganga.sitetest.UKPANDA.muon.
> > 20090728.1.ANALY_R
> > ALPP.13.AANT._00097.root does not exist
> >
> > We've also had a reasonable number of successful runs, and
> so far it
> > looks like the set of WNs that have run jobs successfully does not
> > intersect with those that have failed jobs -- though it's
> probably too
> > early to be drawing conclusions here.
> >
> > Has anyone any idea what could be causing the lack of this output
> > file, or where we could try looking for a problem, please?
> >
> > Thanks,
> > Rob
> > --
> > Scanned by iCritical.
>
>
> Hi
>
> If anyone has installed the latest gLite WN code please be
> warned it breaks ATLAS athena code by including a duffer old
> version of the python logging module and sticking it in the
> PYTHONPATH.
>
> I independently discovered it and then found an existing savannah bug:
>
> https://gus.fzk.de/ws/ticket_info.php?ticket=50148
>
> Glasgow was affected (slightly) and Oxford (almost
> completely). I have applied a work around in the pilot
> factory for now, as well as removed the offending
> grid-cm-client-wn RPMs from Glasgow.
>
> I think you'd be well advised to hold off on upgrades until
> the patched RPMs are released next week.
>
> Graeme
>
--
Scanned by iCritical.
|