Unfortunately I can't do an update as this is production environment.

What is weird is that if I preform ldapsearch from the external to the internal server (Provider) with exact same properties and are configured in syncrepl configuration I see the updated data correctly, therefore the  problem is not writing access rights or something like that. And the data is correctly updated when Consumer is started after the database of the Consumer is deleted.


On Fri, Jan 3, 2014 at 9:34 AM, Dieter Klünter <dieter@dkluenter.de> wrote:
Am Fri, 3 Jan 2014 08:56:27 +0100
schrieb Robert Mach <mach.robert@gmail.com>:

>        Hello,
> I have troubles using openldap replication in Centos 6.4. and Centos
> 5.3 I have two server with following version of openldap from centos
> repository configures as multimaster replication (internal servers):
>
> openldap-servers-2.4.23-32.el6_4.1.x86_64
> openldap-clients-2.4.23-32.el6_4.1.x86_64
> openldap-2.4.23-32.el6_4.1.x86_64
>
> Then I have one external server with following products:
> openldap-servers-2.3.43-12.el5_5.3
> openldap-clients-2.3.43-12.el5_5.3
> openldap-2.3.43-12.el5_5.3
>
> Then two internal servers are configured as multi-master replication
> and everything is working fine. The external server is configures as
> slave replication from one of those internal servers using the
> following configuration:
>
> CODE: SELECT ALL
> database        bdb
> suffix          "o=test,c=cz"
> rootdn          "cn=Manager,o=test,c=cz"
> directory       /var/local/ldap/test.cz
>
> index entryCSN,entryUUID                eq
> index objectClass                       eq,pres
> index ou,cn,mail                        eq,pres,sub
>
> syncrepl rid=132
>   provider=ldaps://xxxxx.xxx.xxx
>   type=refreshOnly
>   interval=00:00:01:00
>   searchbase="o=test,c=cz"
>   filter="(!(objectclass=simpleSecurityObject))"
>   scope=sub
>   updatedn="cn=SyncMaster,o=test,c=cz"
>   bindmethod=sasl
>   saslmech=external
>
> attrs="objectClass,cn,sn,ou,o,mail,serialNumber,userCertificate,cACertificate,certificateRevocationList,authorityRevocationList,entrustPolicyCertificate,crossCertificatePair"
>   schemachecking=on
>
>
> There is a problem with replication from internal server to external.
> If I delete the database of external server and start the consumer,
> everything is correctly replicated from the provider (internal
> server) to the consumer. Therefore I assume, that the replication is
> configured correctly. But if the servers are running for a while and
> changes are made on the Provider (internal server) some data are not
> replicated to the consumer. More precisely the contextCSN of the root
> of the tree of consumer (external server) is updated and is  the same
> as on the Provider(internal server), but some entries lower in the
> three are not replicated correctly, both the entry entryCSN and the
> entry data itself are not updated on the Consumer.
>
> Do you have any idea where could be the problem?

Just update to 2.4.38 and read the CHANGES file.

-Dieter

--
Dieter Klünter | Systemberatung
http://dkluenter.de
GPG Key ID:DA147B05
53°37'09,95"N
10°08'02,42"E