[rrd-users] Consolidation Question

[Mulvaney, Patrick]

Hi,

I am using PNP4Nagios that uses rrdtool 1.4.4 on CentOS 5.5 Linux and I am interested in min/max data points rather than averages.

This may be a rather silly question but when I run the following, which appears to be analogous to the query PNP4Nagios uses internally:-
rrdtool xport --maxrows=10 --start 1293926400 --end 1294920000 DEF:ds1=PING.rrd:1:MIN XPORT:ds1:"Ping"

I get :-

<?xml version="1.0" encoding="ISO-8859-1"?>

<xport>
  <meta>
    <start>1293948000</start>
    <step>108000</step>
    <end>1295028000</end>
    <rows>11</rows>
    <columns>1</columns>
    <legend>
      <entry>Ping</entry>
    </legend>
  </meta>
  <data>
    <row><t>1293948000</t><v>NaN</v></row>
    <row><t>1294056000</t><v>2.6568033333e+00</v></row>
    <row><t>1294164000</t><v>2.6110066667e+00</v></row>
    <row><t>1294272000</t><v>2.6707200000e+00</v></row>
    <row><t>1294380000</t><v>2.7942700000e+00</v></row>
    <row><t>1294488000</t><v>2.6270000000e+00</v></row>
    <row><t>1294596000</t><v>2.5832666667e+00</v></row>
    <row><t>1294704000</t><v>2.6612700000e+00</v></row>
    <row><t>1294812000</t><v>2.6822666667e+00</v></row>
    <row><t>1294920000</t><v>NaN</v></row>
    <row><t>1295028000</t><v>NaN</v></row>
  </data>
</xport>

However if I increase the resolution and filter it for points bellow the min value from above I find values.

rrdtool xport --maxrows=3257 --start 1293926400 --end 1294920000 DEF:ds1=PING.rrd:1:MIN XPORT:ds1:"Ping" |grep 'e+00'|grep '<v>2.[01234]'
    <row><t>1294759200</t><v>2.4431000000e+00</v></row>
    <row><t>1294759800</t><v>2.4431000000e+00</v></row>
    <row><t>1294761600</t><v>2.3702033333e+00</v></row>
    <row><t>1294762200</t><v>2.3317700000e+00</v></row>
    <row><t>1294762800</t><v>2.3317700000e+00</v></row>
    <row><t>1294765800</t><v>2.3038400000e+00</v></row>
    <row><t>1294766400</t><v>2.3947933333e+00</v></row>
    <row><t>1294767600</t><v>2.4536666667e+00</v></row>

I also have a similar problem with MAX which is hiding spikes I wish to report on.

Its probably just me needing pointing in the right direction. However I am seeing spikes in PNP4Nagios's graphs and data being smoothed out which when you want to report on them a bit of an issue.

Thanks

Paddy