[Orca-users] RE: How many orca installations

[Allen Eastwood]

> Well, in the MRTG setup I have now, one machine ssh's to the 
> others every 5 minutes to collect the data, there isn't 
> anything installed on the remote hosts. But then I don't have 
> all hosts setup yet and I do wonder how much load this will 
> create on the MRTG server and on the network when it's 
> querying 60 hosts every 5 minutes about 8-10 different 
> performance criteria. What I like is that there are no 
> dependencies that bind the monitored hosts. They don't mount 
> anything, transfer anything or otherwise do anything that 
> could go wrong and affect their functioning. 
> What looks nice about orca is the way it displays the data 
> according to host or data type, so you can compare in 
> different ways (between hosts, between performance criteria).

The point being that you still have something on the remote hosts
collecting the data, even if it's sar.  And if you're collecting sar
data every 5 minutes, that's using a lot more disk than Orca/SE.

The good thing about the way I have Orca running is that the collection
continues to run even if my NFS or web server were to go down.  So
mounting issues only crop up with looking at the data...if they do come
up.  Once that issue is resolved, I still have complete data.

Personally, it's my opinion that I'd rather have the graph generation
done on each client rather than a central host.  That spreads the load
and I don't have to worry about scaling my processing host.  The perl
scripts that generate the graphs don't impact system performance to a
noticeable degree.  I also use a cron job to update periodically rather
than have the process run continually.
 
> How heavy is SE toolkit? Does it install cleanly? These are 
> in large part productive systems where I can't risk negative 
> effects on the application. They are also managed by a large 
> number of people and it's difficult to ensure homogeneous 
> setups. How heavy is the graph generation?

The SE toolkit is pretty small.  And it has some neat tools of it's own
that I find useful from time to time.  Impact on the system is negible.
Obviously any thing you put on the system will have some sort of impact,
I find that this is much less than other products and even on our
busiest boxes not something you notice.

Our systms are very standardized.  We use jumpstart for everything with
a standard build...makes admining the boxes much easier.  So, the
Orca/SE is just one package that's installed in /opt.  I package them
together, but you don't have to.  SE installs as a standard package by
itself.
 
> Yes, that sounds like a good recommendation. Then I guess 
> it's rather up to me whether the installation is clean. :-)  
> I think I'll try this on some test systems. Does it take a 
> while to get all this setup? Or do the sample config files 
> pretty much do what you need?

The sample/cfg files pretty much give you everythign you need and they
are well documented.  I edit the orcallator.cfg to remove the web server
stats as we don't generally need them.  About the only gotchya in older
version has been systems with large (hundreds) numbers of disks...you
might need to adjust the # of columns in the orcallator.se.  And that
was with 0.26a, I haven't gotten around to updating yet.

-A