[Exim] Exim panics on delivery attempt

Running exim 3.02 on Solaris 2.5.1

Following entry was added to the panic log

1999-09-13 15:31:46 queue run: process 13677 crashed with signal 11 while delive
ring 11QX7k-0003YM-00

whenever exim attempted to deliver this particular message.

Attempting to deliver manually with "exim -d9 -M 11QX7k-0003YM-
00" produces the panic message

1999-09-14 09:57:50 11QX7k-0003YM-00 SEGV while reading "R:compuserver.com" from
dbm file: record assumed not to exist
1999-09-14 09:57:50 11QX7k-0003YM-00 SEGV while reading "R:mikeaziz@compuserver.
com" from dbm file: record assumed not to exist

suggesting something is wrong with with one of the databases or the
dbm support, and indeed "exim_dumpdb /var/spool/exim retry" also
produces a segmentation error.

I've read the FAQ warning of possible problems with dbm library
installation (e.g. mixups with gdbm include files). How can I check
this out? Certainly /usr/include/ndbm.h is full of Sun comments and
does not mention GNU - although at some time I think GDBM has
been installed on the system as there is a gdbm.h file in
/usr/local/include. What else should I look at?

Also, since the retry database seems to be corrupted in some way
(other messages are now producing panic messages) is it safe to
delete it?

Thanks, John








John Holman
QMW Computing Services
Tel: 0171-975 5323