On Mon, Mar 21, 2005 at 03:25:16PM +0100 or thereabouts, Piotr Siwczak wrote:
> Maybe this helps in diagnosing the problem:
>
> when I ran
> globus-job-run ce.egee.man.poznan.pl:2119/jobmanager-lcgpbs -q short
> /bin/hostname
>
> I got:
> /opt/globus/bin/globus-gass-cache: line 3:
> /opt/globus/libexec/globus-script-initializer: No such file or directory
> /opt/globus/bin/globus-gass-cache: line 3:
> /opt/globus/libexec/globus-script-initializer: No such file or directory
> submit-helper script running on host S5-4GB-081 gave error: could not add
> entry in the local gass cache for stdout
>
>
> and in mom logs:
> 03/21/2005 15:13:09;0008; pbs_mom;Req;del_files;cannot stat
> globus-cache-export.K8R06G.gpg
>
> I don't know what it means. Could not find error description anywhere on
> the net.
> BTW: I have /opt/globus/libexec/globus-script-initializer present in the
> required place:
>
> ls -l /opt/globus/libexec/globus-script-initializer
> -rwxr-xr-x 1 root root 1519 Mar 21 12:33
> /opt/globus/libexec/globus-script-initializer
>
> I don't have the script submit-helper present on my WN. Should it be
> there?
The submit helper is on the CE and transfered in with the job that
the job manager creates.
I guess you must have ran
/opt/globus/sbin/globus-initilization.sh
other wise it would be more violent error that you would get.
Steve
>
> Cheers,
> Piotr
>
> Cheers,
> Piotr
>
>
> --
> Piotr Siwczak <[log in to unmask]>
> System Administrator
>
> Poznan Supercomputing and Networking Center
> Supercomputing Department
>
> On Mon, 21 Mar 2005, Maarten Litmaath, CERN wrote:
>
> >On Mon, 21 Mar 2005, Piotr Siwczak wrote:
> >
> >>Hi,
> >>
> >>Thank You for hints.
> >>
> >>I found a few issues that could be sources of my troubles.
> >>
> >>I am having my /home filesystem exported to WNs by NFS. I encountered a
> >>few post regarding "lcgpbs" jobmanager and NFS - some say, that it's
> >>impossible to make them work together. BTW. The "fork" jobmanager works
> >>fine.
> >>
> >>Can this be the reason for my trouble?
> >
> >AFAIK, the "lcgpbs" jobmanager should work fine with NFS.
> >It does not _need_ NFS (it was invented to allow WNs to use their own FS).
> >
> >>On Fri, 18 Mar 2005, Maarten Litmaath wrote:
> >>
> >>>Piotr Siwczak wrote:
> >>>
> >>>>Hi,
> >>>>
> >>>>I can see jobs are submitted to my site, but each one quits with
> >>>>"aborted"
> >>>>state. In globus-gatekeeper logs I find:
> >>>>
> >>>>
> >>>>Failed reading length 0
> >>>>GSS authentication failure
> >>>> globus_gss_assist token :3: read failure: Connection closed
> >>>>Failure: GSS failed Major:01090000 Minor:00000000 Token:00000003
> >>>>
> >>>>Failure: GSS failed Major:01090000 Minor:00000000 Token:00000003
> >>>>
> >>>>Can You help?
> >>>
> >>>For the ID of an aborted job run this:
> >>>
> >>> edg-job-get-logging-info -v 1 $job_ID
> >>>
> >>>The final reason will be "retrycount hit", but look at the earlier
> >>>errors.
> >>>
> >>>Next the Wiki FAQ for job submission problems may tell you what to check:
> >>>
> >>> http://goc.grid.sinica.edu.tw/gocwiki/SiteProblemsFollowUpFaq
> >>>
> >>>BTW, I noticed your site GIIS is down:
> >>>
> >>>$ ldapsearch -x -H ldap://ce.egee.man.poznan.pl:2135 -b
> >>>mds-vo-name=egee.man.poznan.pl,o=grid
> >>>ldap_bind: Can't contact LDAP server
> >>>
> >>
> >
--
Steve Traylen
[log in to unmask]
http://www.gridpp.ac.uk/
|