<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Hi Scott;<br>
<br>
Your approach is absolutely correct -- define multiple instances of the
probe, probably with the 'offset' option, e.g.<br>
<br>
+FPing<br>
offset = 0%<br>
{default fping options}<br>
++FPing1<br>
offset = 25%<br>
++FPing2<br>
offset = 50%<br>
++FPing3<br>
offset = 75%<br>
<br>
Then, depending on the logical structure of your host tree, assign the
probes to hosts, either randomly or by tree, or whatever method you
choose, e.g.:<br>
<br>
+servers<br>
probe = FPing<br>
++server1<br>
++server2<br>
+switches<br>
probe = FPing1<br>
++switch1<br>
++switch2<br>
+routers<br>
probe = FPing2<br>
{...}<br>
<br>
As long as 'concurrentprobes = yes' in your config (which is the
default, but specifying it never hurts), each probe instance will run
concurrently. You just have to be careful not to define so many
concurrent probes that you saturate your network links and skew the
results. :)<br>
<br>
Cheers,<br>
- Peter<br>
<br>
<br>
Scott Moseman wrote:
<blockquote
cite="mid:342d47870711141258y10e98e57w3917be40cfa05a61@mail.gmail.com"
type="cite">
<pre wrap="">It appears my "unreachable" rate is less than 4%. So getting multiple
probes configured appears to be a higher priority than cleaning the
data.
Thanks,
Scott
On Nov 14, 2007 2:34 PM, Scott Moseman <a class="moz-txt-link-rfc2396E" href="mailto:scmoseman@gmail.com"><scmoseman@gmail.com></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Right now I'm using a single FPing probe for all of my devices. We
have a lot of them, and many do not reply, but a "clean up" of the
data is not really an available option at this time. We have a step
of 300 secs and it's beginning to occasionally take more than 300 secs
to complete. In this instance, can I setup multiple FPing probes
(e.g. FPing1, FPing2, etc) and verify concurrentprobes is "yes" (which
I believe is the default)? Is this going to do what I'm thinking it
will, which is running several FPing probes in parallel?
Our config file is automatically generated from a mgmt system. Could
I have the mgmt system output the config with random (or sequentially
changing) probe=FPing* entries? Or is there a better methodology for
distributing the probes among the devices?
Thanks,
Scott
</pre>
</blockquote>
<pre wrap=""><!---->
_______________________________________________
smokeping-users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:smokeping-users@lists.oetiker.ch">smokeping-users@lists.oetiker.ch</a>
<a class="moz-txt-link-freetext" href="https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users">https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users</a>
</pre>
</blockquote>
<br>
</body>
</html>