« 2006-10-28 Thunderbird 1.5.0.8 builds | Main | 2006-11-04 Thunderbird 2.0 builds »
Odometer stopped working
Yes, I know, the odometer has stopped working, and I think it has something to do with the server side controls. mozillaZine.org is hosted over at Oregon University, and they changed certain configurations over at their side sometime this month. I've sent an email to the mozillaZine webmaster but I think the process takes time.
In the meantime, just bear with it. Let's hope that somebody can help with this (I'm not very sure what they initially did to enable it to work); to get it up and running when Thunderbird 2.0 comes out in the future.
Edit: One of the mozillaZine admins, Jason Kersey, has just sent me an email regarding the situation. The counter had been extremely resource-heavy, and according to him, "Every time someone visits your site, it would take up another apache slot, and eventually fill everything up. This, combined with the Firefox 2.0 release, was basically taking the site down. Right now, it'll have to remain off until one of us can figure out the bug. "
I have only basic experience (extremely little to the point of insignificance) in Apache. This would have to wait, then. Any Apache professionals to offer their help?
Hope this clears things up.
Filed under 'zz - Others (pre-2008)' | Posted on 31 Oct 2006 (Tue) by Gary Kwong at 11:42 AM. (GMT+8)
Comments
I thought Jason had already contacted you about this. It was disabled because it was the script was using up virtually all of the Apache worker threads. I'll have a look and see if I can figure out a way to make it work without killing the server.
Posted by: Alex Bishop at October 31, 2006 6:02 PM
One consideration to make on this issue is the use of KeepAlives. You can set in Apaches Environment the amount of simultaneous connections to the server. The more connections the more memory that will be consumed by all process'. One idea might be to set the KeepAlive=Off the downside of this is that users connections will be droppped as soon as their request is filled. This will in turn allow traffic to move across the server faster, but at a cost of performance on larger request.
The second option is to leave KeepAlive=On but limit the KeepAlive time to about 5 seconds. This will increase performance considerably on large page loads but will kill off all unused worker process' after 5 seconds allowing the available worker process' to be freed up much faster.
This is just an idea. In theory it sounds good :)
Posted by: Daniel M at November 1, 2006 6:18 AM
