I’m doing some httperf’ing, and when I’m cranking the load high I see
things like this in apache:
[Mon May 07 19:45:18 2007] [debug] proxy_util.c(1724): proxy:
initialized single connection worker 12 in child 8499 for ()
[Mon May 07 19:45:18 2007] [debug] proxy_util.c(1625): proxy: grabbed
scoreboard slot 12 in child 8500 for worker proxy:reverse
[Mon May 07 19:45:18 2007] [debug] proxy_util.c(1644): proxy: worker
proxy:reverse already initialized
[Mon May 07 19:45:18 2007] [debug] proxy_util.c(1724): proxy:
initialized single connection worker 12 in child 8500 for ()
[Mon May 07 19:45:18 2007] [debug] proxy_util.c(1625): proxy: grabbed
scoreboard slot 12 in child 8501 for worker proxy:reverse
[Mon May 07 19:45:18 2007] [debug] proxy_util.c(1644): proxy: worker
proxy:reverse already initialized
[Mon May 07 19:45:18 2007] [debug] proxy_util.c(1724): proxy:
initialized single connection worker 12 in child 8501 for (*)
I’m guessing thats just noise, and can be safely ignored? googling
didnt’ turn up anything helpful.
Also, this:
[Mon May 07 19:45:19 2007] [info] server seems busy, (you may need to
increase StartServers, or Min/MaxSpareServers), spawning 8 children,
there are 0 idle, and 209 total children
This last one always happens with many children spawned, when I start
using a rate of >70 or something like that. So requests were piling
up in apache, which is expected, but should I really adjust the
prefork settings for this? This is what i have right now:
StartServers 10
MinSpareServers 8
MaxSpareServers 30
ServerLimit 256
MaxClients 256
MaxRequestsPerChild 4000
This is apache 2.3, mongrel latest, ruby 185/rails123 stable, hitting
a pretty complex page that is using a lot of fragment caching in
memcached. No DB hits, though, as everything should be sitting in
memcached. The machine is a quad core intel with 4 gig ram, iirc.
thx,