Well that's a lot of questions. I would first suggest installing Opsview and 
just playing around it. They offer fully functional vm images although the 
setup is pretty simple regardless and they have good install guides.
For me coming for the nagios world understanding the contacts and contact 
methods was the most difficult thing for me to get.
In nagios contacts were in contactgroups and contactgroups were associated with 
hosts and services. With Opsview contacts are associated with hostgroups and 
services can have their own their own notification options or inherit from host.

I love the hostgroup hierarchy features of Opsview, for complex networks you 
can so nested logical grouping of items whereas with nagios you had a rather 
flat listing of hostgroups.
Keep in mind Opsview Master server uses nagios NDO database so IO bottlenecks 
can become a problem a lot quicker with the MySQL dependancies than you would 
have seen with a generic nagios system.

For my company we have started out with distributed monitoring using reverse 
ssh model since we are monitoring many sites across the internet.
Migrate from nagios 2?? I helped one of my customers move from 2 nagios 2.x 
servers to Opsview.  We tested all the custom plugins first, exported all the 
hosts and host dependancies from their nagios servers.
Opsview can import some information from nagios 2.x but in our case we 
programmatically added hosts calling the Opsview API and made heavy use of host 
templates to build up the Opsview system.

"It sounds like the only benefit of Master/Slave Servers are failover?" IMHO 
the master/slave setup gives you more scalability as you can spread the 
processing of checks over n+1 systems.
Also if you are crossing geographic boundaries your polls are occurring close 
to the source they are polling, but you can control and view what is being 
monitored from a central source.

"Both scenarios suggest a separate DB Server. Is this really necessary when 
running on VMware"  You do not want contention of resources between the 
database and the application. Even in a master with multiple slave servers the 
master server has a constant stream of plugin data being returned to it.
I currently have about 500 hosts and 1700 services and  we are planning on 
moving the master server to a faster ESX server and splitting the database to 
it's own server.  In our case we know we are going to be seeing a good bit of 
growth with monitoring so we have to be prepared.

I also have a customer who has 1400 hosts and 2700 services housed on a ESX 
server and I expect the system to start smoking any day  now because the 
hardware just isn't powerful enough.

So virtualization can work but you need fast hardware and you may have to deal 
with time drift issues with 2.6 kernel and ESX.

Overall it's worth it but just knowing nagios only gets you part of the way up 
to speed.

Good luck!


James Whittington
VC3, Inc.






From: [email protected] 
[mailto:[email protected]] On Behalf Of Cook, Garry
Sent: Tuesday, February 16, 2010 4:38 PM
To: [email protected]
Subject: [opsview-users] New User Seeks Advice

I'm about to embark on an Opsview setup for a large company. I've been using 
Nagios since it was NetSaint, so I'm guessing I'll know my way around the 
inside of Opsview. However, I do have a few questions after looking over most 
of the documentation.

First, some background. I haven't upgraded Nagios in quite a while, I'm still 
running v.2.11, monitoring about 600 hosts and 1700 services. Most of these are 
routers and switches with a few WAPs and servers thrown in. I expect that the 
total number of devices will grow to about 800 by year end.
I also run MRTG on a separate server monitoring about 200 of these devices. 
This may grow to about 300, but I expect that Opsview will replace this 
installation as well.
I have two Datacenters, one in Colorado and another in New York, although 
currently the monitoring servers live in the Colorado DC. The servers are all 
Ubuntu running on VMware ESX. Other apps in use include NeDi, SmokePing, 
Weathermap, and two proprietary Netflow apps.

According to the Opsview docs, it seems that distributed monitoring is 
recommended for anything over 250 devices. There is also a recommended hardware 
spec for maximum scalability. So, questions that I have right now:


1.       Should go with the Distributed Monitoring or Maximum Scalability model?

2.       Should I attempt to migrate from Nagios 2, or would it be best to 
build from scratch?

3.       It sounds like the only benefit of Master/Slave Servers are failover? 
So far I've read that you cannot run active/active. Have I just not yet found 
the documentation on load distribution?

4.       Both scenarios suggest a separate DB Server. Is this really necessary 
when running on VMware, since I can just grow the server resources when 
necessary?

Finally, please feel free to offer any other advice that you think might be 
helpful before I dive into this. All input is greatly appreciated.
Thanks!
Garry

________________________________
NOTICE: This e-mail and any files transmitted with it are the property of 
ARCADIS U.S., Inc. and its affiliates. All rights, including without limitation 
copyright, are reserved. The proprietary information contained in this e-mail 
message, and any files transmitted with it, is intended for the use of the 
recipient(s) named above. If the reader of this e-mail is not the intended 
recipient, you are hereby notified that you have received this e-mail in error 
and that any review, distribution or copying of this e-mail or any files 
transmitted with it is strictly prohibited. If you have received this e-mail in 
error, please notify the sender immediately and delete the original message and 
any files transmitted. The unauthorized use of this e-mail or any files 
transmitted with it is prohibited and disclaimed by ARCADIS U.S., Inc. and its 
affiliates. Nothing herein is intended to constitute the offering or 
performance of services where otherwise restricted by law.
_______________________________________________
Opsview-users mailing list
[email protected]
http://lists.opsview.org/lists/listinfo/opsview-users

Reply via email to