On 26 Jan 2001, Cees de Groot wrote:
I've seen Comanche eating CPU on Linux boxen twice now - the first time I thought it was because I had lots of changesets in the image that could be the cause, but my second installation was from a clean image and had the same behavior. It's not much, maybe ~20%, but on an Internet server it's a waste of cycles (especially when said server is doing some unimportant background processing, like hosting 8,000 domains ;-)).
Both installations where Squeak 2.8, Linux 2.2.x on a multi-processor box, with the VM attached to an Xvnc server. Anything that I'm doing wrong with this setup?
I saw this too - same setup as yours (2.2.x, smp, Sq2.8, VNC). It went away some time ago, now CPU usage is less than 1%. Unfortunately, I do not know what I did to cause that ...
Anyway, I just upgraded to RedHat 7.0, and it's still fine. I'm seeing peaks in regular intervals (every 5 minutes or so) and I haven't figured out where these come from
-- Bert
Anyway, I just upgraded to RedHat 7.0, and it's still fine. I'm seeing peaks in regular intervals (every 5 minutes or so) and I haven't figured out where these come from
I noticed that requests make Comanche jump up pretty quickly, but when there are no connections, it goes down to 1% or less. I guess it just has to compete with all the other processes.
Peace and Luck!
Je77
squeak-dev@lists.squeakfoundation.org