replication issue

[Dmitriy Kirhlarov <dimma@higis.ru>]

Hi, list

I have a problem with replication -- numbers of objects on provider and 
consumer not identical.

My configuration:
- two servers (ldap1 & ldap2) configured in mirror mode with carp 
failover. It's work fine.
- also, I have 9 eximx nodes with local ldap servers, configured, as 
consumers for replicate domain routing table from provider.

eximx nodes get replica from shared ip's (i.e. connection possible to 
any from providers)

activity ~ 1-2 changes in minute.

numbers of objects ~ 120 000.
numbers of read on every eximx up to 50 per second

testcase:
1. I deleting db on eximx node and get fresh replica
2. waiting 24 hours
3. check numbers of objects on ldap1 and eximx node.
4. Today, on eximx, I get more objects than ldap1. Yesterday -- eximx 
has less objects, than ldap1.

Could somebody recommend me how to debug this issue or change some 
parameters?

Config on provider:
serverID {1,2} #unique for ldap1 and ldap2
limits dn.exact="cn=root,ou=mail-routes"
        time.soft=unlimited time.hard=unlimited
        size.soft=unlimited size.hard=unlimited
database        hdb
suffix          ou=mail-routes
rootdn          cn=root,ou=mail-routes
rootpw          XXX
directory       /var/db/openldap-data/mail-routes
checkpoint      32 8

overlay         syncprov
syncprov-checkpoint     100 10
syncprov-sessionlog     1000
syncprov-reloadhint TRUE

syncrepl        rid=003
                        provider=ldap://ldap1
                        type=refreshAndPersist
                        retry="60 +"
                        searchbase="ou=mail-routes"
                        schemachecking=on
                        bindmethod=simple
                        binddn="cn=root,ou=mail-routes"
                        credentials=XXX
                        starttls=yes
                        tls_reqcert=allow

syncrepl        rid=004
                        provider=ldap://ldap2
                        type=refreshAndPersist
                        retry="60 +"
                        searchbase="ou=mail-routes"
                        schemachecking=on
                        bindmethod=simple
                        binddn="cn=root,ou=mail-routes"
                        credentials=XXX
                        starttls=yes
                        tls_reqcert=allow

mirrormode      on

index   objectClass,dc  eq
index   entryUUID eq


Config on consumers:
limits dn.exact="cn=owner,ou=mail-routes"
        time.soft=unlimited time.hard=unlimited
        size.soft=unlimited size.hard=unlimited
database        hdb
suffix          ou=mail-routes
rootdn          cn=root,ou=mail-routes
rootpw          ""

directory       /opt/db/openldap/mail-routes
checkpoint      32 8

syncrepl        rid=0
                        provider=ldap://ldap
                        searchbase="ou=mail-routes"
                        type=refreshAndPersist
                        bindmethod=simple
                        binddn="cn=root,ou=mail-routes"
                        credentials=XXX
                        schemachecking=on
                        retry="60 +"
                        starttls=yes
                        tls_reqcert=allow

index           objectClass,dc  eq
index           entryUUID eq


--
WBR
Dmitriy