Anyone using this on OpenBSd


#1

I have just run across Nagios in another forum…
OpenBSD has packages already compiled for
version nagios-2.0b4p0 available…
haven’t gone through the docs completely…
but was wondering if anybody was using this
on OpenBSD…

thx


#2

I use it on Solaris and other OS’s, and from what I"ve seen, BSD is just another linux. Personally, I would install from source only. The nagios docs are for the source install, and with your RPM install, the files will be located all over the place. Unlike the source install that puts all files in the same place. i.e. /usr/local/nagios. Aside from the init script and startup scripts in /etc/rc.d/init.d/ and /etc/rc.d/rcx.d

PS, I strongly suggest you install from source. You will then know exactly where all your files are located, will be able to use the nagios docs more easily, and 95% of the people who use nagios, will be able to assist you quicker, should you need it.
Edited Sun Jan 01 2006, 04:21PM ]


#3

I use Nagios on OpenBSD. It is a little bit tricky for the first time you set up your system, but then it is still reliable.
I build it from source, because the prebuild packages are an earlier version.
The advantage of the prebuild packages on OpenBSD is that you can use a “chrooted” apache and your system is more secure. (Only if you need it)


#4

From what I see about chroot and apache, the only thing you would do different is to install the nagios source to /chroot/httpd/usr/local/nagios instead of /usr/local/nagios. Plus the many other steps to setup the chrooted apache of course.


#5

Thx for the responses…
I don’t have to use a chroot env… but it seems to be only
a few changes… at least in the other packages that I have set
up in a chroot env on my BSD box…

oh… for the record… BSD isn’t Linux… :frowning:

I will go through the docs… I have a project at work that I am looking to use Nagios for…
I have already set up something called “Penemo”… perl network monitor…

It does OK… but from the looks of Nagios, my boss might like this
much better…

my post on another forum which lead me here


#6

The Nagios Package uses differnet directories under OpenBSD. All Under /var/www/
But it is easier for documentation and for other persons to do some maintenance on nagios if you use the standard environment. So finally I “un-chrooted” apache to use the default nagios installation-path.
But there are still a few hints you have to care about…


#7

Kewl…
What kind of issues did you run into… ??
I will probably do the same thing… the chroot isn’t really
necessary in my environment…


#8

Look here:
nagios.org/faqs/viewfaq.php? … desc=false
You must specify the path to the gd library!!

Un-chroot apache you have to use the “U”-Option in the startup script.

Don’t forget to make root own the check_icmp and user nagios is allowed to execute, otherwise every ping check will fail.

User apache doesn’t exit. In openBSD: www

And I have to reconvert some *.gd2 images from nagiosexchange. The statusmap couldn’t show it correctly. But I think this is not an OpenBSD issue, it was a library problem.

After that everything works fine!


#9

SINA…
Much thanks for the pointers…
I have a box here at the house to test this on…
small network… but enough to get more familiar
with Nagios…
oh… and Happy New Year… !!!

thx


#10

:wink: Thanks for the wishes!! Also Happy New Year!
B.T.W.
A few packages are still necessary!
I don’t remember all of them but:
gd-1.8.3, mysql-client (if you are going to use mysql-checks, png, net-snmp,jpeg, …


#11

Update…
I used the packages from OpenBSD, the packages pull in
any dependancies needed… like
gd-1.8.3


#12

Update:
I set up Nagios at work…
I have about 50 ping-poll devices…
Everything is working correctly…

Question…
With this many devices…
are there macros to set up additional
"services"???
I would hate to have to type all those hosts in again when
adding another service…


#13

if you create the same service with the same options for every host you can just add all hostnames in the host line. You can even use *

Luca


#14

I’m working with group names!
You can build groups and then add the services only to the groups.
E.G.: I ping every device, so I added a service ping and then added only the group names to the service.
But I think if you have special services, then it is better to work with the names. This gives you a little bit more control. E.G.: Timeouts, special string answers, etc …


#15

LUCA…
I guess I don’t understand the host line…
are you talking about the hosts.cfg???
or an external file like a /etc/hosts format??

Edit:
I have about 70 hosts that are all ping-poll currently…
I get constant "criticals’ but the host is still up…
Tried replacing the check_ping with check_icmp
which works with the command line… but doesn’t update
the web monitoring page…

thx
Entries for


#16

[quote=“SINA”]I’m working with group names!
You can build groups and then add the services only to the groups.
E.G.: I ping every device, so I added a service ping and then added only the group names to the service.
But I think if you have special services, then it is better to work with the names. This gives you a little bit more control. E.G.: Timeouts, special string answers, etc …[/quote]

SINA… so what do you have in your hosts.cfg file… ???

I have made quite a bit of progress since I started this… but the
more familiar I get with Nagios… the more questions I seem to have…

:cry:


#17

[quote=“thekat”]LUCA…
I guess I don’t understand the host line…
are you talking about the hosts.cfg???
or an external file like a /etc/hosts format??

[/quote]

Yeah, I think luca’s talking about hosts.cfg…if you have a service you want to be exactly the same for a number of hosts, you can have a list of hosts for your host_name option (comma-delimited, of course)…and Sina’s option makes sense, but instead of host_name, I imagine it’d be someting like group_name.

Another option is to define a template for each service that applies to a number of hosts. That way, all you need to have for each service entry would be just a few lines:

define service{
use <template_name>
host_name
register 1
}

Just be sure that when you’re defining your templates, you set register to 0, so Nagios doesn’t think it’s an actual service you want checked. :smiley:

Edited Sun Jan 08 2006, 10:10PM ]


#18

looks like I have more reading to do… :slight_smile:

much tanks for the responses…
Edited Sat Jan 07 2006, 02:23AM ]


#19

Yes, I forget to mention, that I used the hostgroup_names option for the service check.
For your warning problem: Try to figure out how your response time under normal circumstances are. Then change the default values with the -w -c -t, that would help. If you have no reaction with check_icmp then doublecheck the user-rights of your user nagios! check_icmp has to owned by root and nagios has the x right on it. Maybe that’s your problem.
In my host.cfg are only hosts ;). I have onother *.cfg for groups and one for services and one for all templates. It’s better for changes.
I grouped all similar hosts together and then I looked for equal checks, like e.g. open ports. B.t.w. all my hosts do ping checks and this is the reason why I grouped all together to one big group. But one host can be a member in one or more groups. e.g. ALL,WINDOWS,DOMAINCONTROLLER, etc…
But this is only one possible solution.
Try to read : Timesaving tips for object definitions in the official doku.

B.t.w, I’m male! SINA stands for Secure Internetwork Networking Architecture — It’s a VPN Product, so maybe a strange Nick. :wink:

Hope this helps
Edited Sun Jan 08 2006, 08:40AM ]


#20

Sorry it took me a while to get back… was doing more
testing…

my hosts.cfg file is rather large… so far 69 hosts… (well large
when it came to typing in all those hosts… :slight_smile: )
the services.cfg is where I made some changes… following YALL’s advice… :slight_smile:
my services.cfg is much more manageable… thank you…

I am using the default “check_ping” command out of the minimal.cfg…
I will look at the timesaving tips again…

An interesting note…
My first install of Nagios was a 3.2Ghz box OpenBSD 3.8
with 1 NIC…
I got a lot of “criticals” the box is monitoring hosts on the same VLAN but the traffic has to go the router and back to get to that network,
I added a second NIC (on the target network) without changing anything in Nagios
the criticals went away completely…
I installed nagios on my production box (OpenBSD 3.8-stable)
1.4Ghz 256Meg RAM with 2 NICs (one in each network)
the criticals are back…
one of the events
01-11-2006 21:22:56] SERVICE ALERT: hero;PING;CRITICAL;SOFT;1;(Return code of 139 is out of bounds)

SINA, I will work on changing the default values… and get back to the group…

again… much thanks for the response…

rk