How do I clone maintenance mode record ?

I think maybe the server keeps the disabled status in memory, so simply
copying files just isn't going to do it.  This should do the trick:
===== server/ext/getblues.sh =====
      #!/bin/bash

      . /home/xymon/server/etc/hobbitserver.cfg

      $BB localhost "hobbitdboard color=blue
fields=hostname,testname,disabletime,dismsg" | $SED -e 's/\\/\\\\/g'
===== end of file =====

===== server/ext/putblues.sh =====
       #!/bin/bash

       . /home/xymon/server/etc/hobbitserver.cfg

       cat $1 | while read line
       do
          OFS="$IFS"
          IFS="|"
          set $line
          IFS="$OFS"
          x=`echo "$4" | $SED -e 's/\\\n/\n/g'`
          $BB localhost "disable $1.$2 $3 $x"
       done
===== end of file =====

1) on the primary, whenever the other files are copied, run this and copy
the file:

      server/ext/getblues.sh > server/tmp/blues.out

2) whenever the secondary needs to take over, have it run the other script:

     server/etc/putblues.sh server/tmp/blues.out

That should load up the disabled status messages into the secondary.

NB: I've only tested this on a single server, making a slight change to the
messages in the blues.out file before reloading it.  The change loads up
just fine, so it would *probably* work on a different server.  YMMV  :)

Ralph Mitchell


On Thu, Jul 9, 2009 at 1:34 PM, T.J. Yang <user-8e841282cda5@xymon.invalid> wrote:

From: user-c80045f511e8@xymon.invalid
To: user-ae9b8668bcde@xymon.invalid
Date: Thu, 9 Jul 2009 13:22:13 -0500
Subject: RE: [hobbit] How do I clone maintenance mode record ?

I have a separate hobbit-alerts.cfg file for the secondary. It simply
monitors the primary, and if the primary goes down, runs a script to copy
the primaries -alerts file into place.
Then it has a recovery script to turn itself back off.

we use cron job to pull selected bb-host*,hobbit-alert*,hobbit-data and
others from hobbit1 to hobbit2 using rsync.
and with hobbit server side module to check health of hobbit1, if hobbit1
is bad hobbit2 will enable its bbpage worker module to assume the failed
hobbit1 alert service.

Two issues remains.

1. maintenance records are not failed over yet.thus hobbit2 send out false
alerts.
2. rrd files on hobbit1 need to be patched with rrd data from hobbit2 for
hobbit1's down period.


tj

The primary rdists the pertinent files over nightly.


Paul Root IM/MNS Infrastructure
-----Original Message-----
From: Alan Sparks [mailto:user-8f2174fd8b66@xymon.invalid]
Sent: Thursday, July 09, 2009 12:10 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] How do I clone maintenance mode record ?

T.J. Yang wrote:

I am rsyncing configuration files from hobbit1( primary) to hobbit2

(standby) server.

But without maintenance mode records of hobbit1 copy over, hobbit2 will
generate alerts for hosts that had been put into maintenance mode already.

Currently I tried to rsync "data/disabled" directory and restart hobbit2
server but the maintenace web page still show less hosts on hobbit2.

Anyone know where is the maintenance records store at beside

"data/disabled" ?

I recall, from a code browse some time ago, that the alerts module used
the acknowledge.log in deciding whether alerts had been sent -- you may
need to copy that over. I thought it odd, having that in the logs/
directory but used as a data file - but you might look at that.