AW: Nagios as a replacement for HP OpenView

[Mohr James]

Thanks for the detailed reply.

> First HP OV is not one product but 50+, but from
> your questions I think you are referring to HP OV
> Operations (a bundle of NNM, Operations, and
> performance tools). 

My bad. We have OVOU along with OVIS, the OV Service Desk and
Performance Agent. OVIS could easily be replaced with a simply perl
script that checks if certain pages were accessible. 
 
> 1. Network management via SNMP (NNM). HP OV wins big
> time with its ability to auto discover networks,
> map, monitor, track network performance, do basic
> reporting, receive SNMP traps, and generaly manage
> large complex dynamically changing networks. Nagios
> is not nearly as automated and requires other SNMP
> tools to manage traps.

Our networks are fairly static, but the auto discover is a real time
save during the initial setup. However, once that setup is done, things
do not change too much. We do application hosting and our customers
range from 5-20 machines. We have a total of about 120-150 nodes that we
would need to monitor including router, switches, UPSes, etc.

> 2. Server management. Here Nagios gets a lot closer
> to what you get with basic OVO without the smart
> plug ins (SPIs). Nagios is more flexible at hows
> checks are performed to get at info source (snmp,
> local agent, ssh) and since you have source code you
> can probably monitor any brand of system as opposed
> to the 12 or so supported by HPOV.  

We're only running Solaris, Linux and Windows right now, so the support
is less of an issue. However, I didn't think of that point.  


> Nagios' server
> checks are more basic than HPOVs but basic process
> and log file scrubbing plugins are available. 

Can you be a little more specific? What's is missing?

> Three
> addons (NRPE, NSClient, NCSA) are required to come
> up to HP OVOs level of server-agent message
> communication with automatic actions, and you will
> probably have to do some scripting for the automatic
> action parts. If you need the deep application
> monitoring of an OVO SPI, Nagios is not there yet.

Well, as long as I can start anything based on an event, then I say we
are good to go. One "issue" is the annotation in the messages. When
certain events occur we run auto actions that add output as an
annotation. Is there any way that we can do that? Is there any way of
passing "fairly" large chunks of data to the server (e.g. > 2K)? 

What about remote action not related to events? For example, we have
dozens of applications that start processes on the remote machines. Is
there anything built in?

> 3. Performance monitoring/tracking/alerting: Nagios
> is spotty here compared to OVO If you add rrdtool
> and apan, you can get closer, but each performance
> metric must be setup manually in Nagios/apan, and
> the Unix support doesn't seem to be as flexible as
> the Windows support with apan.

What do you mean by "manually"? I imagine that you have do define them
individually in the config files. Is that correct?

Which brings up another point: templates. Other than ftp, scp, etc is
there anything built in to nagios for the distribution of "templates"?

> Overall, I'd say Nagios is great for smaller or
> fairly static IT environments, unless you put some
> effort into automating the Nagios configuration whch
> I think the larger Nagios users have done based on
> some comments on this list. 

We're dealing with ca. 100 agent machines where we would need to
distribute "templates", etc. We create about 5 new templates a week
which we usually need to distribute to just a subset of the machines.
However, "manually" distributing 10 machines is a pain.

> In the bigger picture, not all servers are equal,
> but HP OVO agents cost is  based on server size not
> purpose. 

In my experience it is pretty arbitrary. There are some Sun machines
that are less powerful than Intel machine that are just "servers", but
for these you need to pay "datacenter server" prices. 

> I'd recommend a coexistance strategy until
> such time as you are comfortable that Nagios can do
> it all. For the simple file/print/infrastructure
> servers, redeploy the OVO agent from them to a more
> worthy (complex) environment, and monitor it with
> Nagios. Then put an OVO agent on the Nagios
> management server, and scrube the syslog messages
> with OVOs logfile encapsulator and pass the alert up
> to the central console. 

Now that's not a bad idea for a couple of reasons. First, we don't loose
any functionality during the transition and can see the functionality
side-by-side, thus checking what Nagios misses. Second, this might allow
us to have the two running side-by-side indefinately. The primary
concern is for the low-end customers who just want things at the level
of web-apps or Windows Terminal Server clients. The ratio of agent costs
to the whole maintenance contract is proportionally high compared to
customers using our custom apps. The time save by certain aspects of
OpenView might compensate for the extra work involved in running two
systems.

Hmmmmm.......

Regards,

Jim Mohr







-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id70&alloc_id638&op=click
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null