syncrepl incomplete database problem

[Islam Amer <iamer@open-craft.com>]

Hello everyone,

I have a problem with a simple master/slave syncrepl replication
setup. Everytime I restart the slave it goes back to an empty state
and then syncs about only 200 entries from thousands in the master.
>From then on it gets the update events from the master fine.

I have tried some of the things suggested in this mailing list but
none have worked so far. For example I tried initializing the slave
with a full slapcat of the master after deleting all files from
/var/lib/ldap/ , but when I start the slave it deletes ALL entries and
starts over.

The master is debian etch with 2.3.30-5+etch2 ( critical, not easy to update )

The slave is 2.3.43 RHEL

The slave config is :

syncrepl rid=100 provider=ldaps://xxxxxxxxxxxxxx:636 type=refreshOnly
tls_cert=/var/lib/ldap/ssl/clientcert.pem
tls_key=/var/lib/ldap/ssl/clientkey.pem
tls_cacert=/var/lib/ldap/ssl/cacert.pem interval=00:00:01:00 retry="5
60" searchbase="dc=xxxx" filter="(objectClass=*)" scope=sub
schemachecking=off bindmethod=simple binddn="cn=xxxxx,dc=xxxx"
credentials=xxxxxx

slapd is started with options "-c rid=100,csn=0"

( the real values are obscured for security, sorry )

I switched to refreshOnly because the connection passes through a
strict firewall, and I read that the persistent connection of
refreshAndPersist has problems with that.

Any help is appreciated.
Thanks.