> not that I'm aware of, if you run it on different virtual machines. I never
> tested to run a CE and a WN on the same (physical and virtual) machine,
> though. Maybe some PBS expert is around here?
>
> Did your problem disappear meanwhile? If not, it would be helpful to look into
> the log files, especially /var/spool/pbs/mom_logs on the WN.
Thanks Torsten, sorry to cut into this thread.
i think i could comment more on this issue. further profile directly from
the site CE confirm the root cause could be lie on networking issue, that
we're not for sure if the setup of campus firewall have block out the
outbound access, as all the jobs stuck when downloading the input sandbox,
manualy testing confirm the same problem while local access (edg gridftp
ls and globus url copy) pass normally. we have escalate to site admin and
also informing the network admin, hope to have more input for the changes
done recently. for your information.
i tried bring up another slc3 box (CE as slc4) and configured as mom
yesterday, and again, the torque mom log implying the same error messages.
that i dont think the error related to torque itself, as manualy job
submission could complete normally.
thanks for your input
Br,
J
|