Re: [Exim] exim 4.12 sometimes SIGSEGV

Pàgina inicial
Delete this message
Reply to this message
Autor: Philip Hazel
Data:  
A: avisu
CC: exim-users
Assumpte: Re: [Exim] exim 4.12 sometimes SIGSEGV
On Thu, 13 Mar 2003, avisu wrote:

> some messages are not delivered because auf a signal 11 problem.
>
> Any idea what goes wrong? - it only happens about every 50th msg. And i
> don't have a clue anymore :(


> The part of strace:


If you can get Exim debug output as well, it would be helpful. I can't
tell very much from the strace.

> open("/var/spool/exim/input//18tQcd-000209-00-J",
> O_WRONLY|O_APPEND|O_CREAT, 0640) = 6


That looks like Exim is opening the Journal file in order to record a
successful delivery.

> fcntl64(6, F_GETFD)                     = 0
> fcntl64(6, F_SETFD, FD_CLOEXEC)         = 0
> fchown32(0x6, 0, 0)                     = 0


That's odd. It should be fchowning it to exim:exim, not root. You
haven't set EXIM_USER to root, have you?

> fchmod(6, 0640)                         = 0
> pipe([7, 8])                            = 0
> fcntl64(7, F_SETFL, O_RDONLY|O_NONBLOCK) = 0
> fork()                                  = 11162
> close(8)                                = 0
> getpid()                                = 11160
> wait4(-1, 0xbfffdf70, WNOHANG, NULL)    = 0
> select(8, [7], NULL, NULL, {60, 0})     = 1 (in [7], left {59, 860000})
> --- SIGCHLD (Child exited) ---


That looks like a remote delivery; however the stracing didn't continue
into the child process, so I can't deduce anything else.

Philip

--
Philip Hazel            University of Cambridge Computing Service,
ph10@???      Cambridge, England. Phone: +44 1223 334714.