John,
Is your cluster configured using rsh or ssh between the nodes?
If you are using rsh you have to tell the job manager to use it:
In GLOBUS_LOCATION/setup/globus/ you need to run the following
command
setup-globus-job-manager-pbs --remote-shell=rsh
(http://www.globus.org/toolkit/docs/4.0/admin/docbook/quickstart.html#q-cluster-gram)
Martin
>
> By default it assumes ssh.
> I had properly configured GRAM for GT4.0.5.
>
> Now, trying to integrate with a PBS/MAUI cluster, it looks like work, but
> it doesn't.
>
> The master node of my cluster is the machine pdclab-05.
>
> I run the next command:
> [EMAIL PROTECTED] ~]$ globusrun-ws -Ft PBS -submit -S -f a.rsl
> Delegating user credentials...Done.
> Submitting job...Done.
> Job ID: uuid:adc621f8-4481-11dc-8fe6-00163e6ad4e0
> Termination time: 08/08/2007 01:01 GMT
> Current job state: StageIn
> Current job state: Pending
> Current job state: Active
> Current job state: CleanUp
> Current job state: Done
> Destroying job...Done.
> Cleaning up any delegated credentials...Done.
>
> hmmm!!! no, no...
>
> [EMAIL PROTECTED] ~]$ cat stderr
> Host key verification failed.
> /bin/touch: cannot touch
> `/home/josanabr/.globus/ae1e6610-4481-11dc-9e98-ba518469aa4c/exit.0': No
> such file or directory
> /var/spool/torque/mom_priv/jobs/25.pdclab-0.SC: line 54:
> /home/josanabr/.globus/ae1e6610-4481-11dc-9e98-ba518469aa4c/exit.0: No
> such file or directory
> /var/spool/torque/mom_priv/jobs/25.pdclab-0.SC: line 55: [: too many
> arguments
>
> Now, from pdclab-01 machine, I execute the next command:
>
> [EMAIL PROTECTED] ~]$ globusrun-ws -submit -S -Ft PBS -F
> https://pdclab-05.ece.uprm.edu:8443/wsrf/services/ManagedJobFactoryService
> -f a.rsl
> Delegating user credentials...Done.
> Submitting job...Done.
> Job ID: uuid:f949514a-4481-11dc-bef7-00163e2fd1ce
> Termination time: 08/08/2007 01:03 GMT
> Current job state: StageIn
> Current job state: Pending
> Current job state: Active
> Current job state: CleanUp
> Current job state: Done
> Destroying job...Done.
> Cleaning up any delegated credentials...Done.
>
> I don't have stderr file in pdclab-01, but it is in pdclab-05.
>
> [EMAIL PROTECTED] ~]$ cat stderr
>
> Host key verification failed.
> /bin/touch: cannot touch
> `/home/josanabr/.globus/e9c90e90-4481-11dc-9e98-ba518469aa4c/exit.0': No
> such file or directory
> /var/spool/torque/mom_priv/jobs/26.pdclab-0.SC: line 54:
> /home/josanabr/.globus/e9c90e90-4481-11dc-9e98-ba518469aa4c/exit.0: No
> such file or directory
> /var/spool/torque/mom_priv/jobs/26.pdclab-0.SC: line 55: [: too many
> arguments
>
> I look to the "my_pbs_submit_script" and "my_pbs_job", files created as is
> suggested in
> "http://www.mail-archive.com/[email protected]/msg00553.html";, in its
> content is append to the end of this e-mail.
>
> I think the problem is that the directory is not created by the perl's
> script ${GLOBUS_LOCATION}/lib/perl/Globus/GRAM/JobManager/pbs.pm. I look
> for a line to contains "mkdir" command to create a directory corresponding
> to ${GLOBUS_USER}/.globus/hash_key, and I can't found.
>
> Two questions: Are there something missing in this script related to
> directory creation? If there is, someone has provide a solution for that
> issue?
>
> thanks a lot. regards.
>
> -------------------- my_pbs_submit_script ---------------------------- #!
> /bin/sh
> # PBS batch job script built by Globus job manager
> #
> #PBS -S /bin/sh
> #PBS -m n
> #PBS -o /home/josanabr/stdout
> #PBS -e /home/josanabr/stderr
> #PBS -l nodes=1
> GLOBUS_LOCATION="/opt/gt";
> export GLOBUS_LOCATION;
> X509_CERT_DIR="/etc/grid-security/certificates";
> export X509_CERT_DIR;
> X509_USER_PROXY="";
> export X509_USER_PROXY;
> X509_USER_CERT="";
> export X509_USER_CERT;
> X509_USER_KEY="";
> export X509_USER_KEY;
> HOME="/home/josanabr";
> export HOME;
> LOGNAME="josanabr";
> export LOGNAME;
> SCRATCH_DIRECTORY="/home/josanabr/.globus/scratch";
> export SCRATCH_DIRECTORY;
> JAVA_HOME="/opt/jdk/jre";
> export JAVA_HOME;
> GLOBUS_GRAM_JOB_HANDLE="https://136.145.116.82:8443/wsrf/services/ManagedExecuta
> bleJobService?e9c90e90-4481-11dc-9e98-ba518469aa4c";
> export GLOBUS_GRAM_JOB_HANDLE;
> LD_LIBRARY_PATH=;
> export LD_LIBRARY_PATH;
>
> #Change to directory requested by user
> cd /home/josanabr
>
> hosts=`cat $PBS_NODEFILE`;
> counter=0
> while test $counter -lt 1; do
> for host in $hosts; do
> if test $counter -lt 1; then
> /usr/bin/ssh $host "/bin/sh
> /home/josanabr/.globus/e9c90e90-4481-11d
> c-9e98-ba518469aa4c/scheduler_pbs_cmd_script; echo \$? >
> /home/josanabr/.globus/
> e9c90e90-4481-11dc-9e98-ba518469aa4c/exit.$counter" < /dev/null &
> counter=`expr $counter + 1`
> else
> break
> fi
> done
> done
> wait
>
> counter=0
> exit_code=0
> while test $counter -lt 1; do
> /bin/touch
> /home/josanabr/.globus/e9c90e90-4481-11dc-9e98-ba518469aa4c/exit.
> $counter;
>
> read tmp_exit_code <
> /home/josanabr/.globus/e9c90e90-4481-11dc-9e98-ba518469
> aa4c/exit.$counter
> if [ $exit_code = 0 -a $tmp_exit_code != 0 ]; then
> exit_code=$tmp_exit_code
> fi
> counter=`expr $counter + 1`
> done
>
> exit $exit_code
> ---------------------------------------------------------------
>
> ------------------- my_pbs_job -------------------------
> #!/bin/sh
> cd /home/josanabr
> GLOBUS_LOCATION="/opt/gt";
> export GLOBUS_LOCATION;
> X509_CERT_DIR="/etc/grid-security/certificates";
> export X509_CERT_DIR;
> X509_USER_PROXY="";
> export X509_USER_PROXY;
> X509_USER_CERT="";
> export X509_USER_CERT;
> X509_USER_KEY="";
> export X509_USER_KEY;
> HOME="/home/josanabr";
> export HOME;
> LOGNAME="josanabr";
> export LOGNAME;
> SCRATCH_DIRECTORY="/home/josanabr/.globus/scratch";
> export SCRATCH_DIRECTORY;
> JAVA_HOME="/opt/jdk/jre";
> export JAVA_HOME;
> GLOBUS_GRAM_JOB_HANDLE="https://136.145.116.82:8443/wsrf/services/ManagedExecuta
> bleJobService?e9c90e90-4481-11dc-9e98-ba518469aa4c";
> export GLOBUS_GRAM_JOB_HANDLE;
> LD_LIBRARY_PATH=;
> export LD_LIBRARY_PATH;
>
> ./my_echo "Hello" "World!"
> --------------------------------------------------------------
>
>
> http://ece.uprm.edu/~s047267
> http://del.icio.us/josanabr
> http://blog-grid.blogspot.com
>
>
>
>
