[rrd-users] Strange graph behavior

Petteri Matilainen pmatil at gmail.com
Thu Nov 29 20:53:38 CET 2007


Gary Baluha wrote:
> The other problem with these same graphs is the GPRINT data at the 
> bottom of the graph.  The values that I'm tracking in one particular 
> case is the clock drift of a machine between time updates, and as such 
> the values are usually between 0 and 2000ms.  I'm getting absurdly 
> large values, such as 53934... (a series of non-repeating numbers); 
> numbers so large, they don't even fit on the graph image.  To make 
> matters worse, not every graph is having this issue, and I can see no 
> pattern to the ones that _are_ having the issues.
>
> Any ideas?
I had the same problem with very large numbers with GPRINT. I don't know 
if this helps but in my case, as I'm using an unusual platform (Linksys 
NSLU2 with Debian etch), rrdtool somehow didn't calculate something 
correctly. After searching the 'net a bit I found out that disabling 
optimizations when compiling rrdtool solved the problem (that is, I 
removed -O2 flags from Makefiles). I was having large numbers with every 
graph and it took rrdtool 30 minutes to graph a few images. I'm graphing 
temperature so I also have GAUGE as data type. Good luck.

Petteri Matilainen



More information about the rrd-users mailing list