On 28 Oct 2008, at 13:46, Burke, S (Stephen) wrote:
> Testbed Support for GridPP member institutes
>> [mailto:[log in to unmask]] On Behalf Of Gianfranco Sciacca
>> 10/28 13:18:39 JMI: poll_fast: ******** Failed to find
>> 10/28 13:18:39 JMI: poll_fast: returning -1 = GLOBUS_FAILURE
>> (try Perl scripts)
> I don't think this is a real error, although I might be wrong.
The gram job manager log shows this error over and over as it keeps
trying every 10 seconds. For a successful job it returns 0 =
GLOBUS_SUCCESS, so it looks like a genuine failure after all.
> Have you
> checked the things in this goc wiki entry?
I've checked these and it looks like none of them applies. It looks
like this is now affecting all jobs and SAM tests are staying red. I
wonder whether this may be NFS related as the home directories of the
pool account are setup that way, but we never had a problem with this.
So far I've tried to restart services to no avail and I'm a bit at
loss as to what to try next.
Although I must admit that I'm not sure I understand what the SAM test
show, .i.e. the individual tests are OK (meaning that the output of
the job was retrieved??), but the "js" field stays in "na" state
possibly until the user proxy expires, and then changes to "Error".
Any further ideas?
Dr. Gianfranco Sciacca Tel: +44 (0)20 7679 3044
Dept of Physics and Astronomy Internal: 33044
University College London D15 - Physics Building
London WC1E 6BT