kolab_bootstrap -b with auth troubles

Mark Pierce mark.pierce at telefonica.net
Tue May 18 10:23:57 CEST 2004


Hello Dennis, Hi Martin,

I am in the process of installing Kolab from the Mandrake 10.0 RPMs. 
Open package is a good idea for sharing development but for production, 
I personally would rather reduce the complexity to a single system. You 
could just imagine if every app came out with its own system 
environment... All the same Kolab looks promising, if just judging by 
its components.

I'm getting the same messages as you are from "kolab_bootstrap -b"  But 
I will add that slapd() sends the key messages to /var/log/ldap/ 
ldap.log and thus also to /var/log/messages which repeat each time LDAP 
is started from "service ldap start":

____________________________________________
May 18 10:13:04 karaduen ldap: slapd startup succeeded
May 18 10:13:04 karaduen slapd[7453]: ldbm: ==> unable to initialize 
mutex: Function not implemented
May 18 10:13:04 karaduen slapd[7453]: ldbm: ==> process-private: unable 
to initialize environment lock: Function not implemented
May 18 10:13:04 karaduen slapd[7453]: ldbm_initialize_env(): FATAL error 
in dbEnv->open() : Function not implemented (38)
May 18 10:13:04 karaduen su(pam_unix)[7464]: session opened for user 
ldap by (uid=0)
May 18 10:13:04 karaduen su(pam_unix)[7464]: session closed for user ldap
_________________________________________________


As far as I've been able to ascertain the problem lies in the 
/etc/openldap/slapd.conf  with the line:

database   ldbm

My original ldap configuration uses "database bdm" and starts properly. 
There are several sites which explain the use of multiple databases 
within LDAP out there amongst them:

http://www.tldp.org/HOWTO/LDAP-HOWTO/ldbmdirect.html
http://yolinux.com/TUTORIALS/LinuxTutorialLDAP.html#EMAILCLIENTS

But it will take me a while to wade through all that info to come back 
to whatever errors are written by Kolab into the slapd.conf file.

I'll let you know how I progress, but it would be really nice if someone 
who KNOWS to come in and fix things for us. It's a lot easier to work 
backwards from a running system than to try to figure out the errors as 
you go.

Salud,

Mark Pierce
Xerberos




More information about the users mailing list