mod_proxy details : WAS: Invitation to HTTPD commiters in tomcat-dev
Remy Maucherat wrote: Filip Hanik - Dev wrote: ok, there are two very simple memory friendly ways to do sticky load balancing. And as a matter of fact, this is how some hardware loadbalancers do it. 1. Set a cookie on the clients machine - no server memory to hold a map 2. If the client doesn't accept cookies, do a simple sticky load balancing based on the IP of the client request. Again, no memory map needed. The current jvmRoute addition to JSESSIONID is not really needed, since it doesn't add that much of a benefit over the two options above. So right then and there, there is one less thing to configure. It's cool to have one less thing to configure, but it seems to me jvmRoute is the most reliable and efficient way of doing stickiness (the cookie way is intrusive, and the IP way is highly inaccurate). Well it seems the discussion advance quickly and on the right direction, a true ASF members colaboration. I made some benchs yesterday on my laptop between : - TC 3.3.2/Coyote - Apache 2.0.49 alone (simple html file) - Apache 2.0.49 + jk 1.2.6 + TC 3.3.2/jk2 - Apache 2.0.49 + jk 1.2.6 + 2 * TC 3.3.2/jk2 - Apache 2.0.49 + mod_proxy + TC 3.3.2 (Coyote 1.1). I'll redo them today on a faster machines since the results where a little too random but the benchs raise some questions : - I'm using ab (ApacheBench) and wonder if the -k (keep alive) if HTTP keep-alive is really used ? - Did mod_proxy keep a connection cache ? - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: mod_proxy details : WAS: Invitation to HTTPD commiters in tomcat-dev
Henri Gomez wrote: Remy Maucherat wrote: Filip Hanik - Dev wrote: ok, there are two very simple memory friendly ways to do sticky load balancing. And as a matter of fact, this is how some hardware loadbalancers do it. 1. Set a cookie on the clients machine - no server memory to hold a map 2. If the client doesn't accept cookies, do a simple sticky load balancing based on the IP of the client request. Again, no memory map needed. The current jvmRoute addition to JSESSIONID is not really needed, since it doesn't add that much of a benefit over the two options above. So right then and there, there is one less thing to configure. It's cool to have one less thing to configure, but it seems to me jvmRoute is the most reliable and efficient way of doing stickiness (the cookie way is intrusive, and the IP way is highly inaccurate). Well it seems the discussion advance quickly and on the right direction, a true ASF members colaboration. I made some benchs yesterday on my laptop between : - TC 3.3.2/Coyote - Apache 2.0.49 alone (simple html file) - Apache 2.0.49 + jk 1.2.6 + TC 3.3.2/jk2 - Apache 2.0.49 + jk 1.2.6 + 2 * TC 3.3.2/jk2 - Apache 2.0.49 + mod_proxy + TC 3.3.2 (Coyote 1.1). I'll redo them today on a faster machines since the results where a little too random but the benchs raise some questions : - I'm using ab (ApacheBench) and wonder if the -k (keep alive) if HTTP keep-alive is really used ? Why? Remember to run ab on a separate (fast) machine otherwise the results are random ;-) - Did mod_proxy keep a connection cache ? It does not close the socket to the proxy when using HTTP/1.1 and Connection: is not close. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: mod_proxy details : WAS: Invitation to HTTPD commiters in tomcat-dev
Henri Gomez wrote: I made some benchs yesterday on my laptop between : - TC 3.3.2/Coyote - Apache 2.0.49 alone (simple html file) - Apache 2.0.49 + jk 1.2.6 + TC 3.3.2/jk2 - Apache 2.0.49 + jk 1.2.6 + 2 * TC 3.3.2/jk2 - Apache 2.0.49 + mod_proxy + TC 3.3.2 (Coyote 1.1). I'll redo them today on a faster machines since the results where a little too random but the benchs raise some questions : - I'm using ab (ApacheBench) and wonder if the -k (keep alive) if HTTP keep-alive is really used ? I do my testing with -k usually, and with something like -c 20, to get an average load level (I'm using that on Cygwin, so high concurrency does't quite work in localhost). It seems like keepalive is being used often, and this factors out some of the network stack overhead when I'm profiling stuff. Rémy - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: mod_proxy details : WAS: Invitation to HTTPD commiters in tomcat-dev
Henri Gomez wrote: - I'm using ab (ApacheBench) and wonder if the -k (keep alive) if HTTP keep-alive is really used ? - Did mod_proxy keep a connection cache ? Proxy's HTTP module will reuse the same connection from previous connections if keepalives are being used, it doesn't keep a connection cache for more than one connection at a time. This behaviour is limited to proxy_http though. There is nothing inside proxy that would prevent proxy_ajp from keeping a connection cache. Regards, Graham -- smime.p7s Description: S/MIME Cryptographic Signature
