Re: Apache segmentation faults

* Thus wrote Bostjan Skufca @ domenca.com: > Hello, > > every now and then I notice in apache logs there were few segmentation faults > (on a daily basis) and all I am stuck with is PID of that process (which is > of course dead by then) and nothing about what it was doing. Is there any way > to figure out what request that apache process was serving when SIGSEGV > occured? Is there any reading about this? > > I believe the request is not logged at all because (I think) every child > writes to log files himself and not through parent. (+ log files usually > provide outgoing bytes value, which is not available in such a situation - if > it was logging through parent) > > Again, does anybody know how could I trace out what is causing this? There should be a core file of the dump (httpd.core) generated from the seg fault. If apache is installed in /usr/local/apache, you should find one at: /usr/local/apache/httpd.core You can use gdb to obtain a backtrace to find out exactly where apache died by issuing something like: gdb /usr/local/apache/bin/httpd /usr/local/apache/httpd.core Then type 'bt' Curt -- Quoth the Raven, "Nevermore."