[Orca-users] Re: orcallator client

[Rusty Carruth]

"joeoza" <joza at doubleclick.net> wrote:
> Hello,
> 
> I am trying to setup just the orcallator portion for my clients.  So 
> far I have installed everything (se toolkit/orca/rrdtool/other needed 
> perl modules).  Is there a short cut?  Can I just install se toolkit 
> and then run configure and a option to make to just install 
> orcallator?  Thanks.  

Hmm.  I thought that was in the docs.  Well, look around the docs
and see, if these random comments don't help.

A lot depends upon how you plan to run it.  As I remember it, there
are 3 'processes' going on: data gathering; storing into RRD files;
and, generating the plots.  (well, nitpick it if you will, its ok,
I know I'm not being entirely accurate there ;-)  (and now you do too ;-)

The data gathering pretty much has to run on the machine you
are gathering data for.

The RRD files can be either on that machine, or elsewhere (e.g. via NFS).

The orca process itself can either run one per monitored machine, OR
on a central host.  (I am doing the 'central host' method).

If you do it like I do, all you need on the client machines is an 'install'
of the data gatherer (se toolkit).  In my case, it was pretty trivial,
I just copied everything out of /opt/RICHPse to the remote machines
(since I don't have root on the remote machines, the 'install' is in
my home dir!).  THen I wrote scripts to start the data gathering (SE)
on each remote machine and started SE going.

Then, I started orca on my 'central host' and watched....  Pretty
easy, I thought.

> Also I am running orca as a daemon right now and don't want to get 
> the mail everytime I ssh the files over from the client.  I remember 
> seeing a post that said to run orca from cron; what are the arguments 
> to orca for this?  Thanks.

Cannot help you there, sorry, I just run it continuously...

rc