[Orca-users] Re: How many orca installations
Blair Zajac
blair at orcaware.com
Tue Oct 30 10:22:50 PST 2001
Allen's solution is interesting for performance reasons and doesn't waste
one host that munches data continously. The only drawback to this approach
is that it doesn't allow easy comparisons of the same type of data from
different hosts, which is always good for seeing what's wrong with one system
compared to another. For example,
http://www.orcaware.com/orca/orca-example/o_httpop_per_s,o_http_per_p5s-daily.html
However, we could design Orca to really be two separate pieces of code, a
piece that reads text files and stores the data in RRD files, and another
piece that takes RRD files and plots them in all the ways that Orca does.
This would allow each host to create its own RRD files and plots, and only
use one host to make the plots from all of the RRD files, which should save
CPU resources considerably.
Additionally, this would let us have orcallator.se write RRD files directly,
but I'm a little reluctant to do this, since
1) RRD files loose resolution over time.
2) If you want to change the consolodation interval and other RRD parameters,
this isn't easy.
Blair
Allen Eastwood wrote:
>
> Darie:
>
> I'm not sure what your point is about running something on every host.
> With any performance collection tool, and I've had to evaluate several,
> you have to run some sort of collection process on each machine for
> which you want data.
>
> So, yes, you will have to install the SE toolkit and Orca on each host.
> If you want, you can have the data files sent to a central host for
> processing. Personally, I prefer to have each client generate its own
> graphs and then I send them to a central web server. We have over 100
> boxes currently running this setup.
>
> To make things easier, I would recommend that you you make a package
> with the SE toolkit and Orca and the configuration you want. We also
> use this and it works great, especially since we have it as part of our
> custom jumpstart.
>
> -A
>
> > -----Original Message-----
> > From: duclosd at post.ch [mailto:duclosd at post.ch]
> > Sent: Tuesday, October 30, 2001 08:13
> > To: orca-users at yahoogroups.com
> > Subject: [orca-users] How many orca installations
> >
> >
> > Hi orca users,
> >
> > I've read the documentation inside-out and I feel rather
> > stupid saying that I cannot figure out for
> > sure whether you need to have orca (or orcallator) running on
> > every machine you want to monitor,
> > or only on the "orca server". It looks like it would have to
> > be installed on every machine and the
> > data would somehow have to be visible to the plot-producing
> > machine, but I can't be sure. At the
> > moment I'm using MRTG and I only need to install anything on
> > one host. It rshs (or sshs) to all
> > monitored hosts to gather the data. If this is not the case
> > for orca, i.e. if I have to have it run on
> > every host, then I won't be able to use it unfortunately.
> >
> > Can anyone confirm whether I have to install SE toolkit on
> > every host? Please e-mail me directly.
> > (I'll join the list if it looks like I can use the tool.)
> >
> > Thank you very much.
> > Freundliche Grüsse
--
Blair Zajac <blair at orcaware.com> - Perl & sysadmin services for hire
Web and OS performance plots - http://www.orcaware.com/orca/
More information about the Orca-users
mailing list