We have several queues, primary are default, fast, medium.
The config file is pretty simple: Unscaled[_]: y
Target[sge_1]: /local/packages/IT/MRTG/bin/qgraph.sh both
MaxBytes[sge_1]: 625
AbsMax[sge_1]: 100000000
ShortLegend[sge_1]: Jobs
YLegend[sge_1]: jobs
Options[sge_1]: absolute, gauge, noinfo, nopercent, growright, integer,
unknaszero
Title[sge_1]: Current Jobs on JCVI GRID
Legend1[sge_1]: Running
Legend2[sge_1]: Queued
Legend3[sge_1]: Peak running
Legend4[sge_1]: Peak queued
LegendI[sge_1]: running
LegendO[sge_1]: queued
Colours[sge_1]: GREEN#00eb0c,BLUE#0000ff,GRAY#AAAAAA,VIOLET#ff00ff
PageTop[sge_1]: <H1>Current Jobs on JCVI GRID</H1>
<TABLE>
<TR><TD>Maintainer:</TD> <TD>rdeal at tigr.org</TD></TR>
<TR><TD>Description:</TD><TD>Number of current SGE Jobs running on the
JCVI SGE GRID</TD></TR>
</TABLE>
routers.cgi*Options[sge_1]: noabsmax nodetail
Target[sge_7]: /local/packages/IT/MRTG/bin/qgraph.sh medium
MaxBytes[sge_7]: 354
AbsMax[sge_7]: 100000000
ShortLegend[sge_7]: Jobs
YLegend[sge_7]: jobs
Options[sge_7]: absolute, gauge, noinfo, nopercent, growright, integer,
unknaszero
Title[sge_7]: Current Jobs in medium.q and fast.q
Legend1[sge_7]: Medium
Legend2[sge_7]: Fast
Legend3[sge_7]: Peak medium
Legend4[sge_7]: Peak fast
LegendI[sge_7]: medium
LegendO[sge_7]: fast
Colours[sge_7]: GREEN#00eb0c,BLUE#0000ff,GRAY#AAAAAA,VIOLET#ff00ff
PageTop[sge_7]: <H1>Current Jobs on JCVI GRID</H1>
<TABLE>
<TR><TD>Maintainer:</TD> <TD>rdeal at tigr.org</TD></TR>
<TR><TD>Description:</TD><TD>Number of current SGE Jobs in the fast.q and
medium.q</TD></TR>
</TABLE>
routers.cgi*Options[sge_7]: noabsmax nodetail
The script is rather simple:
#!/bin/sh
case "$1" in
test )
echo 50
echo 20 ;;
total )
echo $T
echo 0 ;;
running )
echo $R
echo 0 ;;
msc )
T=/usr/local/bin/qstat -q fast.msc.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
R=/usr/local/bin/qstat -q msc.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
echo $R
echo $T ;;
medium )
T=/usr/local/bin/qstat -q fast.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
R=/usr/local/bin/qstat -q medium.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
echo $R
echo $T ;;
fast )
T=/usr/local/bin/qstat -q default.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
R=/usr/local/bin/qstat -q fast.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
echo $R
echo $T ;;
desktop )
T=/usr/local/bin/qstat -q desktop.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
R=/usr/local/bin/qstat -q fast.q -s r | /usr/bin/tail +3 | /usr/bin/wc -l
echo $R
echo $T ;;
jtc )
T=ssh wasp "use sgejcvi; qstat -g dt" | /usr/bin/tail +3 | /usr/bin/wc -l
R=ssh wasp "use sgejcvi; qstat -s r" | /usr/bin/tail +3 | /usr/bin/wc -l
echo $R
echo $T ;;
- | both )
T=
/usr/local/bin/qstat -g dt | /usr/bin/tail +3 | /usr/bin/wc -lR=/usr/local/bin/qstat -s r | /usr/bin/tail +3 | /usr/bin/wc -lecho $R echo $T ;; esac
echo foo echo SGE
The two graphs configured above:
From: Martin Flemming <martin.flemming at desy.de> Reply-To: <hobbit at hswn.dk> Date: Fri, 10 Oct 2008 11:46:29 +0200 (CEST) To: <hobbit at hswn.dk> Subject: Re: [hobbit] Sun Grid Engine monitoring
Hmm, i didn't run mrtg ...
How can i do it for sge ?
.. i want to graph e.g jobs in the queue for each node, project or queue ..
... everything i can do :-)
Or is it maybe possible to create graphs with the data of sge.sh script ?
thanks & cheers
martin
On Fri, 10 Oct 2008, michael nemeth wrote:
Ok! well I run mrtg too so I can do that. Also I thought I try creating graphs for the sge.sh when it's implemented. I've done custom graphs before (clearcase). As I said I know little about SGE and the SGE support person knows little about hobbit.
Deal, Richard wrote:
What did you want to graph. I use mrtg to graph global she stats
Sent from a portable system with a small keyboard.
On Oct 9, 2008, at 5:51 PM, "Martin Flemming" <martin.flemming at desy.de> wrote:
Yep, i use it too ... but unfortunatley it's without trending graphs ..
Got anybody a solution for it ?
.. i'm very interested :-)
cheers, martin
On Thu, 9 Oct 2008, rdeal wrote:
It is called sge.sh http://www.deadcat.net/viewfile.php?fileid=920 And we run it under hobbit now. Richard
From: Galen Johnson <Galen.Johnson at sas.com> Reply-To: <hobbit at hswn.dk> Date: Thu, 9 Oct 2008 14:57:10 -0400 To: "hobbit at hswn.dk" <hobbit at hswn.dk> Subject: RE: [hobbit] Sun Grid Engine monitoring
Where on deadcat? I can't find anything that mentions it (a search on grid on Deacat turns up something for the California power grid). Most monitors from deadcat should work with little to no modification for Hobbit. Also, Hobbit will monitor whatever you tell it so the sky's the limit there...or in this case the scripter is the limit.
=G=
-----Original Message----- From: michael nemeth [mailto:michael.nemeth at lmco.com] Sent: Thursday, October 09, 2008 2:46 PM To: hobbit at hswn.dk Subject: [hobbit] Sun Grid Engine monitoring
While I not been involved in the effort yet,we are about to bring as "farm" under Sun Grid Engine in to production. Anyone doing any monitoring like this? I saw a script on deadcat. I really know little about it but said Id check into what monitoring and metrics could be done by hobbit.
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
Gruss
Martin Flemming
Martin Flemming DESY / IT office : Building 2b / 008a Notkestr. 85 phone : 040 - 8998 - 4667 22603 Hamburg mail : martin.flemming at desy.de
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
Gruss
Martin Flemming
Martin Flemming DESY / IT office : Building 2b / 008a Notkestr. 85 phone : 040 - 8998 - 4667 22603 Hamburg mail : martin.flemming at desy.de
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk