[Xymon] Fun with xymon - xynagios and xymondboard

[Betsy Schwartz]

I installed xynagios, which allows dropin execution of nagios plugins in
xymon (and a variant xynagios.yellow )
Since there are a lot of useful nagios plugins out there, this represents a
huge increase in tests we can use off-the-shelf.

Using nagios plugins has an interesting ‘gotcha’. Nagios is event driven ;
if it receives an alert, that alert stays unchanged until either a human
clears it or a new alert arrives.  Xymon however is poll-driven; it expects
alert statuses at regular intervals and turns an alert purple if no updates
occur within the lifetime (set clientside by the “status” variable in the
returned line).  As xynagios is run regularly,  it does the right thing to
update  single alerts. But what happens if a Nagios test sends more than
one sort of alert?

I installed the check_hp_bladechassis plugin, which uses snmp. It sends an
“hp_bladechassis” alert when snmp is up but an “snmp” alert when snmp
fails. So how to keep these from going stale?

What I did was use the ‘xymondboard’ feature to query board status (perl
code below)


If snmp has not logged recently, we know it must no longer be yellow (since
xynagios would have sent a fresh yellow) so set it to green
If snmp is yellow, we know the hp_bladechassis data is stale, so set it to
clear.

   my $boardinfo =
      `$BB $BBDISP "xymondboard host=$blade test=snmp
fields=color,logtime"`;
    ( $color, $logtime ) = split( '\|', $boardinfo );
    if (  ( $epoch - $logtime ) > $STALE )   {
        $bbcmd =
"$BB $BBDISP 'status+2h $blade.snmp green $date SNMP now OK on $blade'";
        system($bbcmd);
    }

    if  ( $color eq "yellow"  )    {
        $bbcmd =
"$BB $BBDISP 'status+2h $blade.hp_bladechassis clear $date SNMP error
unable to
run HP test on  $blade'";
        system($bbcmd);
    }
}
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xymon.com/pipermail/xymon/attachments/20130607/9464527e/attachment.html>