John Doe wrote:
> Only things I found are the hardcoded values in include/net/tcp.h:
I found these tunable parameters: tcp_tw_recycle & tcp_tw_reuse
> Our "issue" is on the LAN side: front servers connecting to the dbs.
> So I wonder if 60s is not too long for the delayed packets problem, when the
> sources and the targets are one gigabit switch away...
Your front end servers should be using connection pooling to go
to the DBs, so there is no delay in having to establish a connection.
Of course connection pooling isn't foolproof I've seen a bunch of
cases where it doesn't work as advertised..
> Hum... I think I just understood why I cap around 14,000 in my tests...
> cat /proc/sys/net/ipv4/ip_local_port_range
> 32768 61000
> (61000-32768)/2 = 14116
> Could that be it?
I don't think so, my settings are the same, and have no problem
getting to 60k+ TIME_WAITs.
nate
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
