Re: OpenNMS server requirements

On 6/17/2010 11:41 AM, Mahoney, John T. wrote:

>>> What are the server requirements to support OpenNMS monitoring about 200
>>> nodes?
>>
>> The usual bottleneck is writing the jrb data to disk, so the number of
>> interfaces, etc. is more important than the number of nodes.
>
> Maybe I am missing the point here, but each of my devices will be monitored on its primary IP/interface.

"Monitoring" and "data collection" are different things.  You can 
monitor a service without storing data or do both (like the response 
time associated with the icmp or http monitor).  Or you can collect data 
without an associated monitor (like snmp traffic/icmp/error counters on 
all ports of a large switch).  If there is a graph associated with a 
value, you are doing data collection.

>>> I am figuring that we will want to run OpenNMS, Postgres, and Tomcat all
>>> on the same server.
>>
>> You'll have much better performance if you can arrange to store the
>> postgres DB, logs, and jrbs on different physical disks to avoid the
>> head contention when these are all written in different places at the
>> same time. Lots of RAM can help a bit with this. Also see
>> http://www.opennms.org/index.php/Performance_tuning
>>
>
> Thanks for the info and link. However, have things changed a lot in the last 4 or 5 years?

RAM and disks are a lot cheaper so it's a lot easier to say 'get more'. 
  And 64-bit boxes that can handle the RAM are more common.

> Back maybe 5 years ago, we had a simple, 2U, RedHat linux server with two hard drives (mirrored) running OpenNMS, PostgreSQL, Tomcat, etc just fine. We are finally ready to stop using Sitescope and go back to OpenNMS but now I am getting concerned about the hardware requirements.
>
> Am I dreaming to think I can ping monitor and do SNMP polling on about 200 different devices/IPs from a single server with two hard drives? On most of these devices, we will probably just send SNMP Traps to OpenNMS and no SNMP polling. (If there is a lot of log data, I can probably store it on our file server/NAS.

You should be able to do the same work on the same box.  I don't think 
the software has gotten any less efficient - but more SNMP oids have 
been added to the default discovery/collection configuration so you 
might have to disable some graphs if you really don't want them. 
Personally, I like having per switchport traffic stats along with the 
host interfaces even though it takes some hardware resources to write 
all that data.

-- 
   Les Mikesell
    [email protected]

------------------------------------------------------------------------------
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-install mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-install