I have a process that cannot be killed by 'kill -9 PID' => kill executes w/o error but the PID is still there (its PPID is 1). The troubling thing is that the process used port 80 before hanging and still keeps a grip on it, i.e. I can't restart apache. I get the same (non)response with /usr/proc/bin/pstop PID
I can only restart apache (2.0.46) on other ports, not on 80. However, when I point an external browser to http://troubled_IP:80 it tries to 'load' forever (10+ min.), and 'netstat -P tcp' shows hostname.80 <details> ESTABLISHED When I stop the external browser 'netstat -P tcp' now shows hostname.80 <same details> CLOSE_WAIT There is a long list of such CLOSE_WAIT entries, just with different <details>. The troubled 'process' is not reported as <defunct> or zombie; there is no 'top' on this sparc // uname -a SunOS tofu 5.6 Generic_105181-17 sun4u sparc SUNW,Ultra-Enterprise =================================================================== How can I kill this process or tell the kernel to release port 80 ? =================================================================== I need to be careful since people are working(*) on this machine around the clock, and a reboot takes about 15 minutes because a lot DB stuff is handled. (*) for that reason I haven't even tried 'ifconfig ifID down' (but when I try this on a test machine connections remains ESTABLISHED even when the interface goes down for a couple of seconds, i.e. I won't do me any good). - Horst Site note: the whole trouble started while testing buggy cgi scripts using the fastcgi module in apache 2.0.46. On working cgi scripts mod_fastcgi is great and can speed up the request-response cycle by a factor of up to 10 :-) _______________________________________________ EuG-LUG mailing list [EMAIL PROTECTED] http://mailman.efn.org/cgi-bin/listinfo/eug-lug
