<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 9/28/2015 2:04 PM, Boldt, David
wrote:<br>
</div>
<blockquote
cite="mid:CAC_ry0bxiwvZ-EsBRxvSydYW5zUBa-QVe1xfv1b5JcPUS1aTMw@mail.gmail.com"
type="cite">
<div dir="ltr">Hello,
<div><br>
</div>
<div>Pardon the newbie questions.</div>
<div><br>
</div>
<div>I can't seem to find any references on the format for "data
messages".</div>
<div>Have tried sending the following, and no rrd file appears
on the server:<br>
</div>
<div><br>
</div>
<div>
<div><font face="monospace, monospace">data
igskahcmasmx01,cr,usgs,gov.msgRate</font></div>
<div><font face="monospace, monospace">messageRate: 1.3980</font></div>
</div>
<div><br>
</div>
<div><br>
</div>
</div>
</blockquote>
<br>
Hi David,<br>
<br>
No worries :)<br>
<br>
Your "data" message syntax is correct, however that by itself isn't
sufficient to generate an RRD file. For that, you generally need to
tell something that listens to the "data" channel that there's
something it should do with "msgRate" data messages. This is
typically done with a configuration in xymonserver.cfg that
indicates a) to listen for the msgRate message, and b) that it's
(probably) in a NameColonValue (NCV) format.<br>
<br>
As a shortcut, and for advanced uses when you want more control over
the specifics of how the RRDs are build, you can use a special
"trends" message<br>
<br>
The mechanics for both of these are detailed at
<a class="moz-txt-link-freetext" href="https://www.xymon.com/help/howtograph.html">https://www.xymon.com/help/howtograph.html</a><br>
<br>
<blockquote
cite="mid:CAC_ry0bxiwvZ-EsBRxvSydYW5zUBa-QVe1xfv1b5JcPUS1aTMw@mail.gmail.com"
type="cite">
<div dir="ltr">
<div><br>
</div>
<div>A second question is on best practices; How should one
launch the data collecting program, with respect to the
definition of the associated test? Below are the current pair
of definitions I am using in clientlaunch.cfg:</div>
<div><br>
</div>
<div>
<div><font face="monospace, monospace">[msgRate]</font></div>
<div><font face="monospace, monospace"> ENVFILE
$XYMONCLIENTHOME/etc/xymonclient.cfg</font></div>
<div><font face="monospace, monospace"> CMD
$XYMONCLIENTHOME/ext/msgRate</font></div>
<div><font face="monospace, monospace"> LOGFILE
$XYMONCLIENTLOGS/msgRate.log</font></div>
<div><font face="monospace, monospace"> INTERVAL 5m</font></div>
<div><font face="monospace, monospace"><br>
</font></div>
<div><font face="monospace, monospace">[msgRate.data]</font></div>
<div><font face="monospace, monospace"> ENVFILE
$XYMONCLIENTHOME/etc/xymonclient.cfg</font></div>
<div><font face="monospace, monospace"> CMD
$XYMONCLIENTHOME/ext/data/msgRate</font></div>
<div><font face="monospace, monospace"> LOGFILE
$XYMONCLIENTLOGS/msgRate-data.log</font></div>
<div><font face="monospace, monospace"> INTERVAL 5m</font></div>
</div>
<div><br>
</div>
<div><br clear="all">
<div><br>
</div>
<div>I'm assuming that I can't run multiple "CMD"s in a single
task definition.</div>
<div>Similarly I'm assuming that I can't reuse the same task
identifier.</div>
</div>
</div>
</blockquote>
<br>
This is correct, each identifier must be unique and is represented
by a single CMD.<br>
<br>
Your script, however, is not simply returning data on STDOUT, but is
probably executing the $XYMON command itself to send a message.
Depending on the data collection you're doing, there's likely little
that would stop you from sending both of the messages from a single
script, though.<br>
<br>
Especially if these are related messages or concept, it's often
simpler to keep collection, evaluation, and other bits relating to a
single logical unit together in one script. Doubly-so if they'd be
running at the same interval anyway.<br>
<br>
<br>
HTH<br>
<br>
-jc<br>
<br>
</body>
</html>