Hi David,
My actual problem was "Job submission failed". Submitting a globus job
says:
[santanu@serv04 test_edg]$ globus-job-run
serv03.hep.phy.cam.ac.uk/jobmanager-lcgcondor -q dteam /bin/hostname
submit-helper script running on host farm030 gave error:
cache_export_dir (/home/dteam005/globus-cache-export.qW58ki) on
gatekeeper did not contain a cache_export_dir.tar archive
and the GOC Wiki page, regarding the 'submit-helper script' leads me
to this conclusion that CLRs are out of date, despite the fact that all
the nodes got latest CA RPMs installed. So, any suggestions, what
should I do now?
Thanks,
Santanu
On May 23, 2005, at 7:04 AM, David Groep wrote:
> Assuming that you run from a UI in Cambridge, the only certificates
> that
> get used in this process are the ones from UKeScience. Since
> (currently) all CRLs are up to date this could have been a local
> update problem.
>
> Anyway, it is likely not related to the errors from SwissSign.
> There seems to be a ongoing problem that the SwissSign web servers get
> overloaded by our CRL requests. This has been reported to SwissSign
> and the SWITCH CA and is currently being investigated.
> (even with only 40 concurrent downloads, about 40% of the CRL requests
> result in a 500-server error or, apparently, a corrupted CRL).
>
> Cheers,
> DavidG.
|