[torqueusers] Hanging TIME_WAIT
tomr at intrinsity.com
Wed Feb 18 00:05:05 MST 2009
We use these settings which seem to minimize the problem:
# release sockets faster because we use a lot of them
net.ipv4.tcp_fin_timeout = 20
# Reuse sockets as fast as possible
net.ipv4.tcp_tw_reuse = 1
net.ipv4.tcp_tw_recycle = 1
Tim Freeman wrote:
> With Torque 2.3.6, we are seeing many connections settle in to the TIME_WAIT
> state and clog up the cluster because of privileged port socket exhaustion.
> Jason Williams reported what looks to be the exact thing last month:
> We're seeing the same netstat output too, the foreign socket is printed as the
> local address.
> Is there anything we can do? Does this imply some misconfiguration?
> torqueusers mailing list
> torqueusers at supercluster.org
More information about the torqueusers