Hi Rayson,<div><br></div><div>Let me first say thanks for OGS, its a super useful tool!</div><div><br></div><div>So, an update....</div><div>I realized that the parameter was load_report_time in the global configuration.</div>
<div>The delay was basically exactly load_report_time, and so I have set it to 0, and the delay is basically gone...</div><div><br></div><div>Rayson, here is my global configuration (qconf -sconf), any comments? Particularly, is it okay to have a value of zero for load_report_time?</div>
<div><br></div><div>$ qconf -sconf</div><div><div>#global:</div><div>execd_spool_dir              /opt/sge6/default/spool</div><div>mailer                       /bin/mail</div><div>xterm                        /usr/bin/X11/xterm</div>
<div>load_sensor                  none</div><div>prolog                       none</div><div>epilog                       none</div><div>shell_start_mode             posix_compliant</div><div>login_shells                 sh,bash,ksh,csh,tcsh</div>
<div>min_uid                      0</div><div>min_gid                      0</div><div>user_lists                   none</div><div>xuser_lists                  none</div><div>projects                     none</div><div>xprojects                    none</div>
<div>enforce_project              false</div><div>enforce_user                 auto</div><div>load_report_time             00:00:00</div><div>max_unheard                  00:05:00</div><div>reschedule_unknown           02:00:00</div>
<div>loglevel                     log_warning</div><div>administrator_mail           <a href="mailto:none@none.edu">none@none.edu</a></div><div>set_token_cmd                none</div><div>pag_cmd                      none</div>
<div>token_extend_time            none</div><div>shepherd_cmd                 none</div><div>qmaster_params               none</div><div>execd_params                 none</div><div>reporting_params             accounting=false reporting=false \</div>
<div>                             flush_time=00:00:15 joblog=false sharelog=00:00:00</div><div>finished_jobs                100</div><div>gid_range                    20000-20100</div><div>qlogin_command               builtin</div>
<div>qlogin_daemon                builtin</div><div>rlogin_command               builtin</div><div>rlogin_daemon                builtin</div><div>rsh_command                  builtin</div><div>rsh_daemon                   builtin</div>
<div>max_aj_instances             2000</div><div>max_aj_tasks                 75000</div><div>max_u_jobs                   0</div><div>max_jobs                     0</div><div>max_advance_reservations     0</div><div>auto_user_oticket            0</div>
<div>auto_user_fshare             0</div><div>auto_user_default_project    none</div><div>auto_user_delete_time        86400</div><div>delegated_file_staging       false</div><div>reprioritize                 0</div><div>
jsv_url                      none</div><div>jsv_allowed_mod              ac,h,i,e,o,j,M,N,p,w</div></div><div><br></div><div><br><div class="gmail_quote">On Wed, Sep 5, 2012 at 12:52 PM, Rayson Ho <span dir="ltr">&lt;<a href="mailto:raysonlogin@gmail.com" target="_blank">raysonlogin@gmail.com</a>&gt;</span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On Wed, Sep 5, 2012 at 1:10 PM, Jesse Lu &lt;<a href="mailto:jesselu@stanford.edu">jesselu@stanford.edu</a>&gt; wrote:<br>

&gt; However, if I run in a parallel environment (e.g. qsub -pe orte ...) then<br>
&gt; there is an approximately 40 sec delay after job completion. That is to say,<br>
&gt; the job has technically finished, although qstat still lists it as running,<br>
&gt; and subsequent jobs are held up. Any ideas?<br>
<br>
</div>That&#39;s fixed in the update release.<br>
<div class="HOEnZb"><div class="h5"><br>
Rayson<br>
<br>
==================================================<br>
Open Grid Scheduler - The Official Open Source Grid Engine<br>
<a href="http://gridscheduler.sourceforge.net/" target="_blank">http://gridscheduler.sourceforge.net/</a><br>
<br>
<br>
&gt;<br>
&gt; Thanks in advance!<br>
&gt;<br>
&gt;<br>
&gt; On Tue, Sep 4, 2012 at 5:33 PM, Rayson Ho &lt;<a href="mailto:raysonlogin@gmail.com">raysonlogin@gmail.com</a>&gt; wrote:<br>
&gt;&gt;<br>
&gt;&gt; That&#39;s the default scheduling time, and if you really want the<br>
&gt;&gt; scheduler to react to your qsub requests ASAP, you can turn on<br>
&gt;&gt; &quot;scheduling-on-demand&quot;:<br>
&gt;&gt;<br>
&gt;&gt; <a href="http://gridscheduler.sourceforge.net/howto/tuning.html" target="_blank">http://gridscheduler.sourceforge.net/howto/tuning.html</a><br>
&gt;&gt;<br>
&gt;&gt; And in OGS/GE 2011.11 u1 p1 (we need a better name), the time it takes<br>
&gt;&gt; to report job done should be reduced.<br>
&gt;&gt;<br>
&gt;&gt; Rayson<br>
&gt;&gt;<br>
&gt;&gt; ==================================================<br>
&gt;&gt; Open Grid Scheduler - The Official Open Source Grid Engine<br>
&gt;&gt; <a href="http://gridscheduler.sourceforge.net/" target="_blank">http://gridscheduler.sourceforge.net/</a><br>
&gt;&gt;<br>
&gt;&gt;<br>
&gt;&gt;<br>
&gt;&gt; On Tue, Sep 4, 2012 at 8:05 PM, Jesse Lu &lt;<a href="mailto:mr.jesselu@gmail.com">mr.jesselu@gmail.com</a>&gt; wrote:<br>
&gt;&gt; &gt; Yes! Exactly.<br>
&gt;&gt; &gt;<br>
&gt;&gt; &gt; -- Jesse<br>
&gt;&gt; &gt; ________________________________<br>
&gt;&gt; &gt; On Sep 4, 2012 4:19 PM, Rayson Ho &lt;<a href="mailto:raysonlogin@gmail.com">raysonlogin@gmail.com</a>&gt; wrote:<br>
&gt;&gt; &gt;<br>
&gt;&gt; &gt; Hi Jesse,<br>
&gt;&gt; &gt;<br>
&gt;&gt; &gt; Are you referring to the scheduling time of Grid Engine??<br>
&gt;&gt; &gt;<br>
&gt;&gt; &gt; Rayson<br>
&gt;&gt; &gt;<br>
&gt;&gt; &gt; ==================================================<br>
&gt;&gt; &gt; Open Grid Scheduler - The Official Open Source Grid Engine<br>
&gt;&gt; &gt; <a href="http://gridscheduler.sourceforge.net/" target="_blank">http://gridscheduler.sourceforge.net/</a><br>
&gt;&gt; &gt;<br>
&gt;&gt; &gt;<br>
&gt;&gt; &gt; On Tue, Sep 4, 2012 at 6:37 PM, Jesse Lu &lt;<a href="mailto:jesselu@stanford.edu">jesselu@stanford.edu</a>&gt; wrote:<br>
&gt;&gt; &gt;&gt; Hi StarCluster users,<br>
&gt;&gt; &gt;&gt;<br>
&gt;&gt; &gt;&gt; I&#39;ve noticed long delays with Sun Grid Engine when submitting jobs and<br>
&gt;&gt; &gt;&gt; especially after job execution. Even running a simple &quot;hostname&quot; job<br>
&gt;&gt; &gt;&gt; takes<br>
&gt;&gt; &gt;&gt; several seconds. Moreover, running an MPI version of &quot;hostname&quot; can<br>
&gt;&gt; &gt;&gt; take 2<br>
&gt;&gt; &gt;&gt; minutes!!<br>
&gt;&gt; &gt;&gt;<br>
&gt;&gt; &gt;&gt; Can someone help me get rid of this delay? Thank you.<br>
&gt;&gt; &gt;&gt;<br>
&gt;&gt; &gt;&gt; Jesse<br>
&gt;&gt; &gt;&gt;<br>
&gt;&gt; &gt;&gt; _______________________________________________<br>
&gt;&gt; &gt;&gt; StarCluster mailing list<br>
&gt;&gt; &gt;&gt; <a href="mailto:StarCluster@mit.edu">StarCluster@mit.edu</a><br>
&gt;&gt; &gt;&gt; <a href="http://mailman.mit.edu/mailman/listinfo/starcluster" target="_blank">http://mailman.mit.edu/mailman/listinfo/starcluster</a><br>
&gt;&gt; &gt;&gt;<br>
&gt;<br>
&gt;<br>
</div></div></blockquote></div><br></div>