*********** REPLY SEPARATOR ***********
On 24.11.2001 at 22:30 Tarjei Huse wrote:
>Have you checked the cyrus-utils.sf.net/faq pages?
Now I have... :o)
>Mika Iisakkila wrote:
>>
>> Georg Altmann wrote:
>> > What does "signaled to death by 11 " mean? Is 11 a pid?
>>
>> No. That means the processes were killed because of signal 11, which
>> means segmentation fault.
ok, that explains some things...
>> You probably have the classic library version clash -- I think RH 7
[..]
>> dynamic linker finds the system libraries, not the ones that you'd
>> hope for.
Ok, I checked the master and imapd binary with ldd. There were in fact old libdb.sos
and an older openssl lib. Though ld had used the right versions I have deleted the old
libs to be shure.
All the other used libs are already new (all built from source).
It seems now, that I have Berkeley DB 3.3.11, openssl 0.9.6b and sasl 1.5.57 cleanly
installed, and, that there are no older (RH) versions lying around. Or would it be
better to use the RH RPMs as recommended in the FAQ?
I configured imapd.conf to use passwd instead of pam for authentication to get rid of
the (possible) ldap clash.
imapd still dies with the segmentation fault error msg.
Could it be some static library causing the trouble? How can I find out against which
libraries imapd gets linked? (Except from searching the whole source for includes...)
>> Another pitfall is the Name Service Switch library. At least in Debian,
>> /etc/nsswitch.conf specifies DB lookups (in addition to files) for
>> services, and the daemons die because the lookup involves DB version 2,
>> and the linker gets confused. Both DB 2 and DB 3 name themselves
>libdb.so,
>> even though the versions are totally incompatible. I chased this problem
>> for a better part of a day.
I found no db entry in the file, shouldn't be the problem.
Thanks for your quick respones!
Regards, Georg
--
Vote against Spam!
http://www.politik-digital.de/spam/de/
The European Coalition Against Unsolicited Commercial Email:
http://www.euro.cauce.org//
