Hi, we're benchmarking our web infrastructure (pound + several nginx) with 300 and more concurrent clients, over 30M hit a day.
Our servers are 1-socket 4-core cpu with 12 GB RAM: Pound starts with 1.2 GB virt memory and increase more and more and allocate all the memory in a day. We try the 2.5 (even with the svc fix found on the mailing list), but nothing happens, so we switch to the 2.6 tree. Our configuration is pretty simple: > ## check backend every X secs: > Alive 5 > > # Main listening ports > ListenHTTP > Address 10.1.20.201 > Port 80 > Client 1000 > > > # Image server > Service > URL ".*.(jpg|gif|png)" > BackEnd > Address 127.0.0.1 > Port 80 > Timeout 30 > End > End > > > Service > HeadRequire "xxx.stardata.it" > Redirect "https://xxx.stardata.it"; > End > > > End > > # SSL section > ListenHTTPS > Address 10.1.20.201 > Port 443 > # Cert "/etc/pound/ssl/www.stardata.lan.pem" > Cert "/etc/pound/ssl/www.stardata.com.pem" > Client 1500 > > Service > BackEnd > Address 127.0.0.1 > Port 80 > Timeout 30 > End > End > End > at 127.0.0.1:80 is listening haproxy. We're using a 5.6 centos distro, x86_64, without google perftools, and we compiled pound with gcc 4.1.2 (we fix configure in order to compile the 2.6e version). Can anyone tell us what is wrong? Thanks in advance, Patrizio -- To unsubscribe send an email with subject unsubscribe to [email protected]. Please contact [email protected] for questions.
