On Mon, Sep 9, 2013 at 2:59 PM, Matthias Hunstock
<matthias.hunstock(a)tu-ilmenau.de> wrote:
Am 09.09.2013 07:14, schrieb Pandu Poluan:
Did my previous email make it into the mailing
list?
Yes it did :)
There is a setting with "reuse" in its name, somewhere in /proc. It is
actually only for tuning but we use it, too. It reduces the timeout of
the waiting connections, which are closed already. The default timeout
is 2 minutes, I think.
Does netstat -ntp print out the PID of a perdition process or a "-" ?
Regards
M.Hunstock
Using 'ss -o state close-wait' indeed returns PIDs of perdition processes.
Currently I did it the brute way: A cronscript running every 30
minutes finds all the PIDs involved and pkill them in one fell swoop.
It's not elegant, but it works...
The sheer number of close-wait connections is a cause of concern for
me, though... in that 30 minute period before the script gets
triggered, the number of close-wait connections reaches a high of
200-something.
Rgds,
--
FdS Pandu E Poluan
~ IT Optimizer ~
• LOPSA Member #15248
• Blog :
http://pepoluan.tumblr.com
• Linked-In :
http://id.linkedin.com/in/pepoluan