[rrd-users] gmetad COUNTER type RRD files contains all zero values!

Guolin Cheng guolin at alexa.com
Thu Dec 11 22:32:50 CET 2008


Hi,

 Any one has encountered similar problem? I've defined COUNTER type
metric with python gmond metric modules. The module works fine and sends
data correctly for other 'GAUGE' type metrics to gmetad.

But for COUNTER type metric, all the valid CDP inside RRD file becomes
zero with only a few exceptions. I suspect that it could be related to
the time issues (step is 15 seconds) of the RRD file, while the gmond
client collecting data every 20 seconds, and thresh-hold is set to 60
seconds.

If the step is a big issues -- since COUNTER type metrics calculate
valid value by subtract previous sample value from current sample value
but the two sample value mostly are the same -- How can I set the 'step'
from metric module?

Another interesting point is that the network bytes in/out, which
normally looks like a 'COUNTER' type metric, is in fact created by
gmetad into a GAUGE type file(s) bytes_in.rrd and bytes_out.rrd, Any one
can explain how these data sequences are collected? 

Thanks.

The following are the rrdfile dump in XML format:

<!-- Round Robin Database Dump --><rrd> <version> 0003 </version>
        <step> 15 </step> <!-- Seconds -->
        <lastupdate> 1229031015 </lastupdate> <!-- 2008-12-11 13:30:15
PST -->

        <ds>
                <name> sum </name>
                <type> COUNTER </type>
                <minimal_heartbeat> 120 </minimal_heartbeat>
                <min> NaN </min>
                <max> NaN </max>

                <!-- PDP Status -->
                <last_ds> 4072050752 </last_ds>
                <value> 0.0000000000e+00 </value>
                <unknown_sec> 0 </unknown_sec>
        </ds>

<!-- Round Robin Archives -->   <rra>
                <cf> AVERAGE </cf>
                <pdp_per_row> 1 </pdp_per_row> <!-- 15 seconds -->

                <params>
                <xff> 5.0000000000e-01 </xff>
                </params>
                <cdp_prep>
                        <ds>
                        <primary_value> 0.0000000000e+00
</primary_value>
                        <secondary_value> 0.0000000000e+00
</secondary_value>
                        <value> NaN </value>
                        <unknown_datapoints> 0 </unknown_datapoints>
                        </ds>
                </cdp_prep>
                <database>
                      <!-- 2008-12-11 13:02:30 PST / 1229029350 -->
<row><v> NaN </v></row>
                        <!-- 2008-12-11 13:02:45 PST / 1229029365 -->
<row><v> NaN </v></row>
                        <!-- 2008-12-11 13:03:00 PST / 1229029380 -->
<row><v> 0.0000000000e+00 </v></row>
                        <!-- 2008-12-11 13:03:15 PST / 1229029395 -->
<row><v> 0.0000000000e+00 </v></row>
                        <!-- 2008-12-11 13:03:30 PST / 1229029410 -->
<row><v> 0.0000000000e+00 </v></row>
                        <!-- 2008-12-11 13:03:45 PST / 1229029425 -->
<row><v> 0.0000000000e+00 </v></row>
                        <!-- 2008-12-11 13:04:00 PST / 1229029440 -->
<row><v> 0.0000000000e+00 </v></row>
                        <!-- 2008-12-11 13:04:15 PST / 1229029455 -->
<row><v> 0.0000000000e+00 </v></row>
                        <!-- 2008-12-11 13:04:30 PST / 1229029470 -->
<row><v> 0.0000000000e+00 </v></row>
....



More information about the rrd-users mailing list