[Orca-users] trying to create my own orcallator...

[Steve Gilbert]

I want to use Orca 0.26 on a Solaris 2.6 machine to graph performance
data from a cluster of Linux machines.  Obviously, there is no orcallator
for Linux, so I'm attempting to create my own using UCD SNMP and some
other tools.  All I want to monitor (for now) are the following stats:

Average # Processes in Run Queue
CPU Usage
Memory Free
Memory Available Swap Space

...I am able to generate those stats easily enough and get them over to
my Orca machine.  I'm looking for a little help in taking this data and
creating a percol-* file.  Right now, I'm thinking something like this:

 timestamp locltime DNnsrkcmdit   uptime  usr%  sys%  1runq  5runq 15runq  swap_avail  freememK
 970038001 00:00:01 wwwwwwwgwww  4460363   0.4   0.6   0.00   0.01   0.01      130504    506560

...my question is about the first four fields and which ones I need and
don't need.  First, if this stuff is already documented somewhere, please
point me towards it.  I couldn't find much while digging through the
distribution.  My questions are:

1. How is that "timestamp" generated?
2. Do I have any need for the "DNnsrkcmdit" status field, and if so, what
   on earth does it all mean?  I looked at orcallator.se, but it's not
   obvious (to me anyway) how those are used.  I'm hoping I can just cut
   it out if it's not needed for my purposes.
3. Do I have any need for the "uptime" considering what I want to monitor?

Thanks for any help/advice anyone can offer.  This is going to be a
challenging project for me.

Steve Gilbert
gilbert at cs.utk.edu

-------------------------- eGroups Sponsor -------------------------~-~>
Get FREE long-distance phone calls on Tellme!
Dial 1-800-555-TELL, say "Phone Booth"
http://click.egroups.com/1/9816/13/_/804404/_/971917838/
---------------------------------------------------------------------_->