Sean,
Any idea on how to handle the seg fault below? It happens nearly
every time I do a graceful restart. If that can be taken care of, I
wonder if this patch might be considered, in experimental/, for the
apache mainline?
Thanks for all your work!
Brock
> everything happens just after graceful
>
> Program terminated with signal 11, Segmentation fault.
> #0 apr_socket_addr_get (sa=0x8315014, which=APR_LOCAL, sock=0x0) at
> sockaddr.c:196
> 196 if (sock->local_interface_unknown ||
> sock->local_port_unknown) {
> (gdb) bt full
> #0 apr_socket_addr_get (sa=0x8315014, which=APR_LOCAL, sock=0x0) at
> sockaddr.c:196
> No locals.
On Thu, Feb 14, 2008 at 6:14 PM, Sean Gabriel Heacock
<[EMAIL PROTECTED]> wrote:
> My apologies for not being very active on the list lately, as usual I
> have too many projects going on and no free time...
>
>
> On Thu, 2008-02-14 at 12:38 +0100, Lazy wrote:
> > from my testing i think it crashes when there are no file descriptors left,
> > after that the serwer is in unresponsive state
>
> Well that would do it. Have you tried raising FD_SETSIZE? On linux
> it's defined in /usr/include/linux/posix_types.h and is 1024 by default.
> If it's still set that low, raise it to something like 8194 and
> recompile apache and peruser.
>
> You can also run "ulimit -n 8192" from apachectl or an init script, but
> I believe you can't set a higher limit than is defined by FD_SETSIZE.
>
> I wonder... maybe we can redefine FD_SETSIZE inside mpm.h? I hate
> making changes to core include files like posix_types.h, since it just
> gets wiped out next time I upgrade glibc.
>
> --
> Sean Gabriel Heacock
> Telana Internet Services
> http://www.telana.com/
>
>
>
> _______________________________________________
> Peruser mailing list
> [email protected]
> http://www.telana.com/mailman/listinfo/peruser
>
_______________________________________________
Peruser mailing list
[email protected]
http://www.telana.com/mailman/listinfo/peruser
