[rrd-users] Re: Possible problem?

[Tobias Weingartner]

On Thursday, July 29, Tobias Oetiker wrote:
> 
> hmmm this looks rather scary ... 
> 
> <!-- 1999-07-28 08:35:00 MDT --> <row><v> 8.5340093594e+02 </v><v>-6.49045672
> 81e+92 </v></row>
> <!-- 1999-07-28 08:40:00 MDT --> <row><v>-6.4904574033e+92 </v><v> 3.53465479
> 02e+03 </v></row>
> 
> I suspect you are seeing an overflow problem ... your database does not set
> either min nor max for the datasouces ... 
> 
> the data display routines of rrd_graph prbably click out at about 10e18 ....
> 
> it would be interesting to research into why the generated graph looks the
> way it does ... anyone ?

Well, adding '-r -l 0' to the generation of the graph, does show a spike at
roughly the same area of time.  I am noticing something else in my graphs.
They are very "spiky".  I'm not sure if this is the usual thing to expect
from "barely loaded" switches...

For a peak, feel free to browse the preliminary cricket tree at:
http://irricana.cs.ualberta.ca/~cricket/grapher.cgi

--Toby.

--
* To unsubscribe from the rrd-users mailing list, send a message with the
  subject: unsubscribe to rrd-users-request at list.ee.ethz.ch