linux - IPv6 vs IPv4 connectivity load -


one of our applications uses cross-platform tcp connectivity layer. layer has connectivity stress-test. test launches 20 client threads , 20 server threads; each client connect/small-data-exchange/close several randomly chosen server threads.

the test done 4 connectivity variants (ipv4 ipv4, ipv6 ipv6, ipv4 dual-mode, , ipv6 dual-mode). passes on everywhere except on one particular 64-bit linux machine 4 machines of running redhat 2.6.18-8.el5. ipv4 connectivity (both ipv4 , dual-mode) passes on this machine these machines, ipv6 can handle tenth of should able to. gets timeout errors, few connection-reset errors.

cpu, memory, descriptors, etc. not issue. i've reviewed network settings on machine, nothing seems have been messed with.

using localhost vs host name changes nothing. (in particular, i'm ruling out faulty network card, fails on loop-back.)

netstat shows nothing unusual. (a lot of sockets in time_wait, that's expected given nature of test.)

i'd appreciate thoughts on issue be.

ipv6 on rhel5.0 cannot handle load. requirements our app moved rhel 5.3.


Comments

Popular posts from this blog

ruby on rails - RuntimeError: Circular dependency detected while autoloading constant - ActiveAdmin.register Role -

c++ - OpenMP unpredictable overhead -

javascript - Wordpress slider, not displayed 100% width -