It is by no means a substitute for the bundled documentation or even basic common sense: you'll be doing all this as root, so if you screw things up, you've been warned.
It makes several assumptions:
You need the following packages/tarballs: the RICHPse* toolkit , including virtual Adrian, and ORCA itself.
Orca is a generally used to describe the data gathering and plotting components, but that's not very accurate. The data gathering and reporting is all done with the Rich Pettit and Adrian Cockcroft tools (the SE toolkit and ANC rules). Orca is an extension of the rules (orcallator.se, descended from percollator.se: see Adrian's book for more on that), which is then used by the RRD toolkit, itself descended from MRTG. The plots are made by orca/rrd and the data is extracted/derived by orcallator.
You also need to plan your installation. You need to know where you want things to end up on the clients and the central server (where the plotting tool actually runs). Here's where the assumptions come into play, but I'm going to show what I did so you can adapt it to your needs.
On each client machine (where the orcallator.se tool runs), you need to install the RICHP and ANC packages, doing ANC last: the others are prerequisities.
By default, these end up in /opt, so make sure you have space there.
/usr/sbin/pkgadd -d . RICHPse RICHPsex ANCrules
Next, set up your NFS stuff. A sample:
orcabox:/usr/local/apache/htdocs/orca - /mnt/orca/htdocs nfs - yes
rw,bg,hard,intr,proto=udp
orcabox:/usr/local/var/orca - /mnt/orca/rrd nfs - yes rw,bg,hard,intr,proto=udp
NB: these must be writable, so make sure you have space for N clients to store their data.
The first mount point is where the actual user-viewable stuff goes. The second is where the orcallator output goes for rrd to work with.
Make the necessary substitutions in the command line below. I plucked these out of my history file, so they actually worked for me.
Now, unroll the orca tarball (0.26 at this writing). Assuming you've read all the docs that come with orca and laid everything out correctly, copy and paste should work. Feel free to use these as examples.
./configure --with-orcallator-dir=/mnt/orca/rrd/orcallator --with-rrd-dir=/mnt/orca/rrd/rrd/orcallator
--with-html-dir=/mnt/orca/htdocs
export PATH=${PATH}:/usr/ccs/bin
make test_modules install_modules install orcallator_run_at_boot
the default installation root is /usr/local/[lib|bin] so if you want to change it, use the --prefix argument to configure. /opt/orca might be a good idea, for example.
Double-check your installation, looking over orcallator.cfg and start_orcallator. The default plots may not be what you need to start out with, but there should be more than enough data to get you going.
If all looks well, ./start_orcallator in your $bindir.
This will get data written out to the NFS mount point you set up. Repeat this as necessary across all the systems you want to orcallate.
On the hub machine, your orcabox in the vfstab entries above, you should now be able to run /path/to/orca /path/to/orcallator.cfg. You can use the -v argument to get verbose output, if you like.