The patch I attached to that bugzilla report does exactly this, except I
used a null output filename:

http://bugzilla.ganglia.info/cgi-bin/bugzilla/attachment.cgi?id=15&action=view

I just looked at the most recent rrdtool documentation:

http://oss.oetiker.ch/rrdtool/doc/rrdgraph.en.html

and it does say that if stdout (-) is used as the output image filename
with the graph command, that no other output is generated.  This
explains why the two print statements are ignored in newer versions of
rrdtool when the ganglia code uses stdout.

Maybe this can be fixed in the cvs/svn source code for the next version
release?  I also have a few other minor patches in bugzilla:

http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=30
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=36

One to add some useful "metrics" back into the web frontend which were
present in much older versions of ganglia and the other one to fix a
minor bug affecting the node display for dead nodes.

~Jason


On Wed, 2006-05-10 at 14:10 -0700, Chuck Simmons wrote:
> Aha!  After reading the rrdgraph man pages on the net and screwing 
> around a bunch, I modified the code in find_limits to look like:
> 
>         $command = RRDTOOL . " graph /tmp/dummygangliagraph.out ".
>             "--start $start --end $end ".
>             "DEF:limits='$rrd_dir/$metricname.rrd':'sum':AVERAGE ".
>             "VDEF:limitsmax=limits,MAXIMUM ".
>             "VDEF:limitsmin=limits,MINIMUM ".
>             "PRINT:limitsmax:%.2lf ".
>             "PRINT:limitsmin:%.2lf";
> 
> The dummy output filename was critical to get any output at all.  Other 
> changes avoid deprecated syntax or make the code match the documentation.
> 
> Thanks, Chuck
> 
> 
> Jason A. Smith wrote:
> 
> >Hi Chuck,
> >
> >This is how it works for me.  The cluster_view.php code sets a min & max
> >value, using the find_limits function in functions.php.  Then the
> >graph.php code uses these values to set the rrdtool upper & lower limits
> >in the graphs.  About a year ago I discovered a bug that might have been
> >related to the version of rrdtool that I was using, which effectively
> >broke the find_limits function.  I reported this to ganglia's bugzilla:
> >
> >http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=37
> >
> >Maybe you are affected by the same problem.
> >
> >~Jason
> >
> >
> >On Wed, 2006-05-10 at 12:10 -0700, Chuck Simmons wrote:
> >  
> >
> >>I want to display multiple graphs with the same upper-limit value.  
> >>Basically, I need to obtain the maximum upper limit from across the 
> >>graphs (or a good estimate thereof) and then use that when displaying 
> >>the graphs.
> >>
> >>For example, on cluster_view.php, we will display one graph for each 
> >>node in a cluster.  The graph displayed is chosen by the user from a 
> >>pull down menu.  At the current times, the graphs are displayed at 
> >>widely different scales, and a human cannot quickly graph at the screen 
> >>to determine which nodes have the highest values.  The human's life 
> >>would be easier if all graphs used the same scale.
> >>
> >>What are the 10 best ways to implement a solution to this problem?
> >>
> >>Thanks, Chuck
> >>
> >>
> >>-------------------------------------------------------
> >>Using Tomcat but need to do more? Need to support web services, security?
> >>Get stuff done quickly with pre-integrated technology to make your job 
> >>easier
> >>Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> >>http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> >>_______________________________________________
> >>Ganglia-developers mailing list
> >>[email protected]
> >>https://lists.sourceforge.net/lists/listinfo/ganglia-developers
> >>
> >>    
> >>
> 
-- 
/------------------------------------------------------------------\
|  Jason A. Smith                          Email:  [EMAIL PROTECTED] |
|  Atlas Computing Facility, Bldg. 510M    Phone:  (631)344-4226   |
|  Brookhaven National Lab, P.O. Box 5000  Fax:    (631)344-7616   |
|  Upton, NY 11973-5000                                            |
\------------------------------------------------------------------/



Reply via email to