On Sat, 20 Apr 2002, Sam Michaels wrote:
> Having this problem....some of the delivery processes crash with signal 11.
> Not sure what to do..
>
> An 'strace' gets this far:
>
> --snip---
> open("/var/spool/exim/db/retry", O_RDONLY) = 7
> fcntl(7, F_SETFD, FD_CLOEXEC) = 0
> fstat(7, {st_mode=S_IFREG|0640, st_size=20480, ...}) = 0
> lseek(7, 0, SEEK_SET) = 0
> read(7, "\0\0\0\0\0\0\0\0\0\0\0\0a\25\6\0\5\0\0\0\0\20\0\0\2\0\0"..., 4096)
> = 4096
> --- SIGSEGV (Segmentation fault) ---
> +++ killed by SIGSEGV +++
> ---snip---
>
> It's been attempting to deliver the messages since Feb 13th....I figured the
> retry times would have expired by now.
>
> Finally gave up and did an -Mrm on them. Any ideas what would cause this?
This looks like a crash inside your DBM library. The most likely cause
is a corrupt DBM file /var/spool/exim/db/retry. Alternatively, you
rebuilt Exim with a different DBM library and it crashes on trying to
read a file in the other format. "Different" in this case also includes
different releases of Berkeley DB.
Try running "exim_dumpdb /var/spool/exim retry" and see if it crashes.
If it does, I suggest you remove the entire contents of the db
directory. This kills all Exim's retry hints. It will then recreate a
new set as time goes on.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
