[op5-users] Merlin

Andreas Ericsson ae at op5.se
Fri Apr 30 11:03:09 CEST 2010


On 04/29/2010 11:46 PM, Tyler Burrows wrote:
> Wondering if I can direct a couple quick questions to you. It seems
> there's a bit of confusion on our end getting the information we need
> to make a decision on merlin/nagios as a NMS solution.
> 

Sure. I'm redirecting this to op5-users as well, so more can benefit
from the answers, and perhaps someone else can chime in with answers
from their own experience as well.

> The real basic requirement is using nagios/merlin to monitor a
> distributed network comprised of many different clients on many
> sites.
> 

Neat. That's exactly why Merlin was brought to life in the first
place. You'll want to use the 'next' branch of the merlin repository
though, as we're keeping 'master' as a stable branch for the 0.6
series of Merlin for now. You can clone it by issuing the command

  git clone git://git.op5.org/nagios/merlin.git

There's no configure script, so just type

  make

to compile it. You'll need libdbi-devel and the standard development
tools for your server to do this.

> We want to be able to poll the networks and move that data back to a
> central server where the hosts from each of those sites would be
> separated into hostgroup. It would be nice if this was done
> automatically from the poller config. Perhaps this is done already?
> 

Yes. Each poller is intended to be responsible for its own hostgroup.
Check the example config for more information regarding this.

Configuration isn't distributed automatically though, so you'll have
to sort that out yourself for the time being. It's scheduled to be
implemented in a "won't cover all cases, but the ones we guess are
most useful" fashion when merlin's distributed functionality is
supposed to be stable, around august this year.

> How does SLA reporting work in this situation? Can we do
> sub-hostgroups for routers, workstations and servers etc?
> 

You can have as many hostgroups as you like. Only the hostgroups that
are mentioned in the merlin configuration will be used to hand out
checks to the pollers.

> If the connection to the main server is cut off, will the poller cue
> that data until the connection is restored?
> 

Not yet, and never indefinitely. There is a binlog api that can and
will be used for this though, but it's not plugged into the networking
code yet.

> Is there any way to do custom nagvis maps per customer based on host
> groups or would we have to go through and code this?
> 

nagvis and its integration into ninja isn't my child, so I don't know
for sure. I *do* know that users are supposed to only see hosts they
are contacts for though, so I suppose you could fix this the standard
way and a little "try and find out" experimentation.

> Thanks in advance for all your time! If there's any config guides
> that answer the above it would be appreciated if you could direct us
> to them!
> 

All the merlin documentation is in the merlin repository and on the
op5 wiki, located at https://wiki.op5.org/merlin:start
Feel free to update the wiki with any useful information or trouble-
shooting tips you come up with.

-- 
Andreas Ericsson                   andreas.ericsson at op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231

Considering the successes of the wars on alcohol, poverty, drugs and
terror, I think we should give some serious thought to declaring war
on peace.


More information about the op5-users mailing list