[rrd-users] Confused about unknown values

[Oliver Gehrke]

I'm trying to use rrdtool to monitor an array of machines on a LAN,
using a script that queries them via SNMP. I've got rrdtool 1.2.6
(Linux/2.6.14) to store and graph the data, so far so good.

Now I've discovered a problem with values supposed to be unknown (channels
not being updated because the corresponding computers are down):
they don't show up as unknown/NaN in the database - instead they
are zero.

My database looks like this (only three channels here, to make it easier):

rrdtool create /var/rrd/traffic/node-in-traffic --start N --step 10 DS:syslab-00-
nicin:COUNTER:100:U:U DS:syslab-01-nicin:COUNTER:100:U:U DS:syslab-02-
nicin:COUNTER:100:U:U RRA:AVERAGE:0.5:1:480 RRA:AVERAGE:0.5:6:2520 
RRA:AVERAGE:0.5:60:2016

and when, let's say, the first one of the computers (DS "syslab-00-nicin") is down,
the update command being issued by my script would look like this:

rrdtool update /var/rrd/traffic/node-in-traffic -t syslab-02-nicin:syslab-01-nicin 
68650972:369163891

As I interpret the documentation, the first column (syslab-00-nicin) should be filled with 
NaN, since no data is ever coming in - but it has zeros instead. This creates problems
particularly if zero cannot intuitively be related to a box being down (for network traffic
it makes sense, for cpu idle% it doesn't really).

Does anyone have any ideas? I've been trying a lot of things and potential workarounds, 
but it seems I got stuck here.

Oliver

--
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