Some common problems and troubleshooting:
<ul>
<li>  qstat -f shows node in E (error) state.

This has happened when qsub used to submit actual binary programs
rather than shell scripts.  Each node the job is submitted to dies and
is set to E state, and the SGE system retries the job a couple of
times causing two node queues to be marked as E each time.

<br>
manual says:
<blockquote>
       If an E(rror) state is displayed for a queue, sge_execd(8)
       on that host was unable to locate the sge_shepherd(8) exe-
       cutable on that host in order to start a job. Please check
       the error logfile of that sge_execd(8) for leads on how to
       resolve the problem. Please enable  the  queue  afterwards
       via the -c option of the qmod(1) command manually.
</blockquote>
in our setup:
<pre>
     grep NODENAME /sge/default/spool/qmaster/messages
</pre>
for hints of failure.<br>
Example:
<pre>
Thu Dec 11 11:32:51 2003|qmaster|cyclades|W|job 3079.1 failed on host node003 general assumedly before job because: can't write script file "job_scripts/3079" wrote only -1 of 893666 bytes: Bad address
Thu Dec 11 11:32:51 2003|qmaster|cyclades|E|queue node003.q marked QERROR as result of job 3079's failure
</pre>

You can clear the E status of all queue nodes with qmod:
<pre>
[root@cyclades default]# qmod -c node*.q
[root@cyclades default]# qstat -f
...
node003.q            BIP   0/2       0.00     glinux    
...
</pre>

</li>
</ul>