[rrd-users] rrd updates give NaN (rrd4j)

[Michele Mazzucco]

Hi all,

I'm using the rrd4j library (https://rrd4j.dev.java.net/), a java
implementation of RRDTool, to monitor my system.
I've already posted the same question on the rrd4j mailing list, however
it's not very "popular" and thus I got no replies :( (and sorry for the
cross-posting).

Premise that my updates can happen very close each other, even every
second, the problem is that all values into the RRD db are set to NaN.
The parameters I've set are 10 as step (but I've tried with the default
value of 300 seconds as well) and 50 as heartbate (but, again, I've
tried with Long.MAX_VALUE as well).

Some more details:
- some datasources are declared as "gauge" while others as "counter",
- all archives keep track of the average (the X-files factor is 0.5, the
number of archive steps is 1 -- 6 in the long-term archive -- and the
number of archive rows is 6 -- 24 in the long term archive)
- some datasources are updated by +/-1 (that is, the last value is
used), while other datasources simply write a value. While I could
understand the NaN for the first case (the first NaN is propagated), I
don't understand why, for datasources which simply write a pair
<timestamp, value>, the result is NaN.

What I think is that the period between two consecutive updates is too
short, and thus rrd4j is not happy :(. Has anybody aver had any
(successful) experience with very-frequent updates (even with RRDTool)?


Thanks in advance,
Michele


--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://lists.ee.ethz.ch/rrd-users
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi