perl, sasl or cyrus problem?

suffocator suffocator at gmx.de
Tue Jun 22 22:30:58 CEST 2004


Hi Denis,

Tuesday, June 22, 2004, 5:55:14 PM, you wrote:

>> >> as written before on this list I have had several problems to get
>> >> Kolab installed. Mostly I get the famous 'no worthy mechs found'-error
>> >> in /var/log/messages. I've seen that there has been a bug report for
>> >> SuSE 8.1, obviously this problem isn't solved, is it?
>> >>
>> >> I've tried several ways to install Kolab (ZFOS and source, both in
>> DC> different
>> >> versions) on Redhat9, Fedora Core1, RHEL3, Debian and Mandrake. For me
>> >> things just worked using Debian 3.0 (after several attempts with
>> >> kolab2) and Mandrake 10 (worked out of the box). I followed the
DC> source-QIM
>> >> precisely - no luck, ZFOS installation works fine - no luck, tips from
>> DC> pokorra.de
>> >> did not work for me too.
>> >>
>> >> So the question is why does this error occur, what am I and others
>> >> doing wrong? Maybe we can split the problem into several pieces to
>> >> test and verify.
>>
>> DC> I have also been through this process and managed with the Redhat
DC> 3.0/ZFOS
>> DC> setup but now even that no longer wants to work for me.
>>
>> DC> How do you install it under Mandrake ? because I can no longer afford
DC> to
>> DC> waste my time trying to get any of the other versions to work and need
DC> a
>> DC> system that I can get working 100% every time & keep up to date.
>>
>> DC> Many thanks
>>
>> DC> Denis
>>
>> I just downloaded MDK10 + Cooker pkgs (~5GB) and solved dependencies
DC> manually
>> with help from rpmseek.com or console output requiring other packages. You
>> have to install normal RPMS as well as Cooker RPMS, it's a bit
>> confusing jumping between cooker and normal RPMS, but it works for me.
DC> Horde
>> webclient installation works too. But there was an authentication
>> issue different from native Kolab, I think you just have to use
>> <username> instead of <username at domain.com> when adding users using the
>> webinterface. Accessing the admin webinterface you have to use
>> https://host.domain.com/kolab instead of ~/admin.
>>
>> Be aware that you have to use several packages that are still in
>> cooker-status.

DC> Thanks for that, I was unable to download as quick as I needed the CD's so I
DC> have just purchased the Mandrake 10.0 Powerpack + and did a server install
DC> chose the "Kolab" server option and it worked 100% so far. And I have 8CD's
DC> 1 DVD & 2 manuals.
DC> (After approx 6 weeks, 5 disto's and over 40 install's of hell the cost was
DC> worth it, so was the 3 hours to go and get the pack)

DC> By the way, Do you know where the ldap data is stored in the Mandrake
DC> version  as I now need to setup a backup routine !

DC> Regards

DC> Denis

Good news, by the way I checked /var/log/messages on mdk10 and indeed there is
'perl: no worthy mechs found' too, but somehow it works anyway... I
think there's some dirty hack concerning sasl/ldap/cyrus.

At the moment I do not have a box with mdk running, but normally all
LDAP-data is stored unter /var/lib/ldap, there might be several
.dbb-files. Backing up LDAP-Data has two or three pitfalls, depending
on the availability you're trying to achieve. Easiest way is to stop
ldap, backup /var/lib/ldap/* and /etc/openldap/* and start again. In
most cases this would work. Regarding HA you will have to a) choose
another backend with online backup-ability or b) replicate your
data using slurpd to a slave-server, stop ldap-service on slave, backup
data on slave, startup ldap-service on slave. In case you have to
restore your data you can copy your backup into /var/lib/ldap. You can
also use slapcat (man slapcat will help) and export your data to a
ldif-file and backup this one. In case of restore you would use
slapadd to restore your database with that ldif-file.


-- 
Best regards,
 Magnus                            mailto:suffocator at gmx.de




More information about the users mailing list