Yes, sorry guys to write this here , but i installed ubuntu (i swear that i only use debian normally) - but this was a reverse proxy with a Dual Xeon System with 12 gig ram and other things, and i needed a brutal new linux to handle this (and i really did not want to use debian testing for this system) (but i already feel now like a beta-tester -- i will never do this again -- but now too later)
The machine handles a reverse caching proxy in front of a webappearance (mpm-worker) , and having 400 hits / second. It appered, that all 2 min i had a funny "proxy error" around, and i almost got mad, as on the testing system i installed b4 (debian) this mistake was not existing. i found out after reading 3 days --- and trying everything, that this was a bug comming from apache 2.0.54 to 2.0.55 AND ! a patch exists ... but bad, not in your repository (btw. the patch is already included in 2.0.58, maybe even sooner, i did not look) a link here http://people.apache.org/~wrowe/httpd-2.0-proto-timeout.patch This patch works direct in source package of 2.0.55-4 maybe you add it, i am sure, some other people who operate Apache as reverse proxy will sleep a little bit more better when you know, that not every 1:2000 website requests a "Bad Gateway error" appears on the user webbrowser. btw, i know, compleate offtopic here, but i get about 2 messages / day about Jul 18 23:08:57 localhost kernel: [877685.871379] APIC error on CPU3: 60(60) i have read, that this is not brutal dramatic until it gets really high, but does anybody have an idea ? kernel is 2.6.15-25-amd64-xeon franz
