Configure Syncrepl with subordinate db

[Petter Solgaard <petter.solgaard@ntnu.no>]

Hi

I'm trying to set up at new ldap-cluster (openldap 2.4.9 on Ubuntu 
8.04). I have splitt the DIT into two dbs, where I have a ou=system 
under the base that is a subordinate db to the top db. Both dbs is on 
the same server. What I am wondering is where I should insert the 
replication information. I tired to put the replication information 
under the top db on the master and under the top db on the slave. What 
happens is that information is replicated between the top dbs, but not 
the subordinates (master to server). I tried adding an additional rid on 
the subordinate db on the slave and this made the replication work for 
the subordniate also. But when I the tested with a reboot of the servers 
the base became corrupt and I had to install everything all over again. 
Could someone confirm that the replication information should be under 
the top db on the provider and under both dbs on the consumer?

***********************
* Slapd.conf provider:          *
***********************
modulepath      /usr/lib/ldap
moduleload      back_hdb
moduleload      back_monitor
moduleload      back_bdb
moduleload      syncprov
moduleload  accesslog

---------SUBORDINATE----
database        hdb

suffix          "ou=system,dc=......,dc=com"
limits          dn.exact="cn=replicator,dc=......,dc=com" size=unlimited 
time=unlimited

subordinate

index   entryUUID,entryCSN       eq

access to *
   by dn="cn=replicator,dc=.....,dc=com" read
   by peername.ip=127.0.0.1 read
   by * none

--------ACCESSLOG-----

database        hdb
rootdn          "cn=accesslog"

overlay syncprov
syncprov-nopresent TRUE
syncprov-reloadhint TRUE
limits dn.exact="cn=replicator,dc=.....,dc=com" time.soft=unlimited 
time.hard=unlimited size.soft=unlimited size.hard=unlimited

index           default eq
index           entryCSN,objectClass,reqEnd,reqResult,reqStart eq

--------TOP-----
database        hdb

suffix          "dc=.....,dc=com"

rootdn          "cn=Manager,dc=....,dc=com"
rootpw          verysecret


### Replikerings-innstillinger ###
overlay glue
overlay syncprov
syncprov-checkpoint 1000 60

# accesslog overlay definitions for primary db
overlay accesslog
logdb cn=accesslog
logops writes
logsuccess TRUE
logpurge 07+00:00 01+00:00

# Let the replicator DN have limitless searches
limits dn.exact="cn=replicator,dc=....,dc=com" time.soft=unlimited 
time.hard=unlimited size.soft=unlimited size.hard=unlimited

index   entryUUID       eq

lastmod         on

access to *
   by peername.ip=127.0.0.1 read
   by ssf=128 users read
   by * none


***********************************
* Slapd.conf consumer:            *
***********************************
# Where the dynamically loaded modules are stored
modulepath      /usr/lib/ldap
moduleload      back_hdb
moduleload      back_monitor
moduleload      back_bdb

-------SUBORDINATE------
database        hdb

suffix          "ou=system,dc=....,dc=com"
limits          dn.exact="cn=replicator,dc=....,dc=com" size=unlimited 
time=unlimited

index   entryUUID                eq

access to *
   by dn="cn=replicator,dc=....,dc=com" read
   by peername.ip=127.0.0.1 read
   by * none

--------TOP-----

database        hdb

# The base of your directory in database #1
suffix          "dc=....,dc=com"

# rootdn directive for specifying a superuser on the database. This is 
needed
# for syncrepl.
rootdn          "cn=Manager,dc=....,dc=com"
rootpw          verysecret


# Replication from master
syncrepl rid=100
 provider="ldap://ldap.****.****.com:389";
 type=refreshAndPersist
 searchbase="dc=.....,dc=com"
 filter="(objectClass=*)"
 scope=sub
 attrs="*"
 schemachecking=on
 bindmethod=simple
 binddn="cn=replicator,dc=.....,dc=com"
 credentials="secret"
 logbase="cn=accesslog"
 logfilter="(&(objectClass=auditWriteObject)(reqResult=0))"
 retry="60 +"
 syncdata=accesslog


# Refer updates to the master
updateref       "ldap://ldap.****.****.com:389";

index   entryUUID       eq

lastmod         on

access to *   
   by peername.ip=127.0.0.1 read
   by ssf=128 users read
   by * none



--
Kind regards
Petter S