Dave Ingram wrote:
On Fri, 11 Oct 2002, Steven Wagner wrote:


I have a G4 tower here on my desk which I *could* use for Jaguar development. However, in my computing environment it doesn't make much sense to pursue it aggressively.


        A very fair response. :)

Unfortunately, now you've got me curious so I'm looking into it a bit. :) It looks like you would mix up some sysctl() calls for the basic system attributes (CPU speed, hostname, kernel version, etc.) with some calls to kvm for the kernel statistics like memory usage and CPU cycles. From that description, feel free to copy large chunks of code from my solaris.c and irix.c implementations if you're looking to just get something to compile (it might work...).

Good grief - I wouldn't have expected it to share very much with Tru64. We have that here on a few machines as well and while it's a bit scary, I still think HP-UX takes the cake for weirdness.

I did say they were "surface similarities."  :)

I can probably find someone here to port it, but I thought I would check first. Last time I checked, we develop Mathematica for Linux (x86, PPC, Alpha), Tru64, AIX, Darwin, Solaris and HP-UX. So I'd imagine I can dig up some Darwin guru to hack away at it. We're working very closely with Apple on this cluster project, and we'll most likely be getting similar deals together with a few other big names. It would be very cool to have all of them running Ganglia!

Personally, I would try to get the input from at least one long-term (does such a thing exist at this point?) Darwin sysadmin and at least one Darwin kernel hacker on what metrics are worth monitoring. I would especially not be shy about getting app-specific, if it's useful for your purposes...

If we get a successful port, I'll post information here and provide a copy of the finished source. Thankfully, I've got some of the right people here salivating enough that this should happen soon.

Shop towels are usually pretty cheap at Costco...



Reply via email to