[GE users] Problems with LAM tight integration
bli at bcgsc.ca
Thu Aug 3 20:46:21 BST 2006
[ The following text is in the "iso-8859-1" character set. ]
[ Your display is set for the "ISO-8859-10" character set. ]
[ Some special characters may be displayed incorrectly. ]
Isn't your job script missing something like:
#$ -pe lam 10
From: slaton [mailto:slaton at berkeley.edu]
Sent: Thu 03/08/2006 12:06
To: users at gridengine.sunsource.net
Subject: Re: [GE users] Problems with LAM tight integration
> > whereas the rsh wrapper generated this (note additional lines not
> > included above):
> > /opt/sge/bin/lx24-amd64/qrsh -V -inherit -n -p 32796 qcn11 exec
> > '/opt/sge/utilbin/lx24-amd64/qrsh_starter'
> > '/opt/sge/default/spool/qcn11/active_jobs/59.1/1.qcn11'
> It looks, like the rsh-wrapper is called with parameters which are
> already processed by an rsh-wrapper and now the final command is covered
> again by an rsh-wrapper. Do you have any additonal directories in your
> PATH, which might use another rsh-wrapper (where you commented out the
> "echo") - any $SGE_ROOT/mpi or so?
There IS an rsh wrapper in /opt/sge/mpi, which is used for mpich parallel
environment. However /opt/sge/mpi is not in PATH; only
/opt/sge/bin/lx24-amd64 and /usr/userstat/bin.
The only other thing i could think of is that in my environment module for
LAM, i set LAMRSH to 'rsh', which is just actually unnecessary since it is
the default (unless compiled with --with-rsh ssh or some such thing). I
removed this definition and it made no difference.
I also tried setting LAMRSH to '/opt/sge/lam_tight_qrsh/rsh' which also
made no difference.
Since this LAM installation is dedicated for SGE tight integration jobs,
do you think it would help to Just for completeness i will also try
recompiling LAM with --with-rsh configured to /opt/sge/lam_tight_qrsh/rsh,
./configure --with-rsh="/opt/sge/lam_tight_qrsh/rsh" [..other opts..]
For what it's worth the test script i'm submitting is very simple:
#$ -N mpihello
echo "using tmpdir $TMPDIR"
/usr/local/lam/7.1.2/sge/pgi/bin/mpirun C ./mpihello
> > i'm perplexed as to how..
> > $rhost $cmd
> > expands to..
> > -n -p 32795 qcn15 exec '/opt/sge/utilbin/lx24-amd64/qrsh_starter'
> > '/opt/sge/default/spool/qcn15/active_jobs/61.1/1.qcn15'
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net
More information about the gridengine-users