I gave this issue some more thought. While the MAX setting on my graphs
stray as I increase the time frame, it has never been a huge concern of
mine, since we manually record the last days max reading. That said,
this issue has got me thinking.
I've never been convinced that the issue has to do with having multiple
data point resolutions, since the only thing that changes is the number
of data point between the recordings. (So the max for a one day period
recorded every 5 minutes will be the same for the max value recorded for
the day.) As stated the graph will perform a consolidation if you try to
display too many data values for the number of pixels in the graph. But
even when I force the graph size to match the data points, the max value
slowly erodes. When I do a fetch for a months worth of data, I notice
that there are a handful of "nan" readings in the middle of the data
set. I suspect that these may be affecting the consolidation function
when graphing. There are a handful of tricks to handling these
situations when generating a graph. The CDEFTutorial on the RRD website
it goes into them.
On Wed, 2007-07-18 at 09:21 -0400, James Roman wrote:
> The only way I know of to troubleshoot actual errors when generating an
> rrd is to manually create a graph using the rrd statement from the
> command line. If you can't do it directly in RRD, then Zenoss won't be
> able to display it anyway.
>
> One way to proceed within Zenoss is to make a dummy graph first, with
> just the standard options. Select your data point. Don't put anything in
> your Custom Command section. Now view that graph. You should now be able
> to see the command that was used to generate that output by using "tail
> -10 /opt/zenoss/log/Z2.log | grep datasource_datapoint.rrd" Copy the
> output from the "DEF" statement up to, but not including the "&drange"
> statement.
>
> Now you have a baseline statement to work with. Since you want to modify
> the DEF statement, create a new graph, but make sure that you don't
> select a data point, otherwise Zenoss will try to include its own DEF
> statement. Simply pasting the original statement should display the same
> graph as was generated before. Proceed by adding one statement at a time
> and verifying that it will still display.
>
> I don't really know if you will be able to mess with the start and end
> statements within the DEF statement,as they are issued as part of the
> command line options. But there is rub, if you can't control the start
> and end time, it means you can't calculate for the proper step value.
>
>
>
> On Tue, 2007-07-17 at 22:34 +0000, agthurber wrote:
> > I am unable to get any of the DEF functions to produce graphs, whenever i
> > try to access the RRD file my graph is broken. I have tried every variation
> > of path to access it but none of them appear to work. It is quite
> > frusterating because there is no debug output to let me know what i am
> > doing wrong, if anything,
> >
> > any pointers on accessing the RRD files with a DEF statement? the path to
> > my RRD dir is
> >
> > /usr/local/zenoss/perf/Devices/{DEVICE NAME}/{DATA POINT}.rrd
> >
> > what path should my DEF statement be using to find my RRD files?
> >
> > ------------------------
> > A. G. Thurber
> >
> >
> >
> >
> > -------------------- m2f --------------------
> >
> > Read this topic online here:
> > http://community.zenoss.com/forums/viewtopic.php?p=8845#8845
> >
> > -------------------- m2f --------------------
> >
> >
> >
> > _______________________________________________
> > zenoss-users mailing list
> > [email protected]
> > http://lists.zenoss.org/mailman/listinfo/zenoss-users
--
James D. Roman
IT Network Administration
Terranet Inc.On contract to:
Science Systems and Applications, Inc.
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
