<div><br><div class="gmail_quote">2010/11/15 Arnau Bria <span dir="ltr">&lt;<a href="mailto:arnaubria@pic.es">arnaubria@pic.es</a>&gt;</span><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">

<div class="im">On Mon, 15 Nov 2010 13:15:57 -0200<br>
Denis Denis wrote:<br>
<br>
&gt; Could you send you maui.cfg?<br>
</div>Sure (I&#39;ve added a couple of node bewteen lines).<br>
<br>
<br>
SERVERHOST              NAME<br>
ADMIN1                  root<br>
ADMIN3                  edginfo rgma edguser monami<br>
ADMINHOST               NAME<br>
RMCFG[base]             TYPE=PBS TIMEOUT=30<br>
SERVERPORT              40559<br>
SERVERMODE              NORMAL<br>
<br>
RMPOLLINTERVAL        00:02:00<br>
LOGFILE               /var/log/maui.log<br>
LOGFILEMAXSIZE        50000000<br>
<br>
IDLEJOBDEPTH  300<br>
#This come from a patch<br>
#<a href="http://www.supercluster.org/pipermail/mauiusers/2009-February/003746.html" target="_blank">http://www.supercluster.org/pipermail/mauiusers/2009-February/003746.html</a><br>
<br>
<br>
<br>
BACKFILLPOLICY          NONE<br>
BACKFILLDEPTH           1<br>
LOGLEVEL                1<br>
<br>
LOGFILEROLLDEPTH        50<br>
<br>
ENABLENEGJOBPRIORITY true<br>
REJECTNEGPRIOJOBS false<br>
<br>
QUEUETIMEWEIGHT         0<br>
<br>
XFACTORWEIGHT           0<br>
<br>
<br>
CREDWEIGHT              1<br>
GROUPWEIGHT             1<br>
USERWEIGHT              1<br>
CLASSWEIGHT             1<br>
<br>
NODEALLOCATIONPOLICY    CPULOAD<br>
<br>
DEFERTIME               00:00:00<br>
<br>
CLASSCFG[long]          MAXPROC=100<br>
CLASSCFG[medium]        MAXPROC=100<br>
GROUPCFG[dteam]         MAXPROC=40 PRIORITY=10<br>
GROUPCFG[dtsgm]         MAXPROC=2 PRIORITY=100000<br>
GROUPCFG[dtprd]         MAXPROC=20 PRIORITY=100000<br>
GROUPCFG[ops]           MAXPROC=20 PRIORITY=100000<br>
GROUPCFG[pilotops]      MAXPROC=20 PRIORITY=100000<br>
USERCFG[arnaubria]      PRIORITY=1000<br>
<br>
SRCFG[picsgm_64]        GROUPLIST=atsgm,sgmcm,lhsgm,masgm,ctasgm,dtsgm,misgm,pasgm,picvosgm,sgmibergrid<br>
SRCFG[picsgm_64]        RESOURCES=PROCS:4<br>
SRCFG[picsgm_64]        PRIORITY=1000<br>
SRCFG[picsgm_64]        HOSTLIST=tditaller021<br>
SRCFG[picsgm_64]        STARTTIME=0:00:00 ENDTIME=24:00:00<br>
SRCFG[picsgm_64]        PERIOD=INFINITY<br>
<br>
FSWEIGHT                1<br>
FSUSERWEIGHT            2<br>
FSGROUPWEIGHT           10<br>
FSQOSWEIGHT             100<br>
<br>
FSDEPTH                 4<br>
FSINTERVAL              12:00:00<br>
FSDECAY                 0.5<br>
FSPOLICY                DEDICATEDPS%<br>
<br>
<br>
<br>
GROUPCFG[masgm]         FSTARGET=10  QDEF=magic MAXPROC=2<br>
GROUPCFG[maprd]         FSTARGET=10  QDEF=magic<br>
GROUPCFG[magic]         FSTARGET=10  QDEF=magic<br>
QOSCFG[magic]           FSTARGET=5.79<br>
[....]<br>
<br>
OTHER QOS CONF<br>
[...]<br>
<br></blockquote><div><br></div><div>what does a diagnose -p report?<div>Is it possible that the jobs which are running before your highest priority job are not being backfilled but having a higher priority instead due to the weights of the other metrics?<br>

</div></div><div>I see that the CREDWEIGHT is set to 1 while QOS for example is set to 100.</div><div><br>Also there are some groups with priority really high ( 100000)</div><div><br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">


<br>
<br>
What are you looking for?<br>
This is my test conf which looks very similar to prod one (except<br>
backfill params)<br>
<div><div></div><div class="h5"><br>
Cheers,<br>
Arnau<br>
_______________________________________________<br>
mauiusers mailing list<br>
<a href="mailto:mauiusers@supercluster.org">mauiusers@supercluster.org</a><br>
<a href="http://www.supercluster.org/mailman/listinfo/mauiusers" target="_blank">http://www.supercluster.org/mailman/listinfo/mauiusers</a><br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>Denis Anjos,<br><a href="http://www.versatushpc.com.br">www.versatushpc.com.br</a><br>
</div>