<p>I'm far from being an expert but have you considered clearing the existing replication agreements and following the instructions for setting up multi-master from scratch? </p>
<div class="gmail_quote">On Jul 13, 2011 5:44 PM, "Roland Schwingel" <<a href="mailto:Roland.Schwingel@onevision.com">Roland.Schwingel@onevision.com</a>> wrote:<br type="attribution">> Hi...<br>> <br>
> Since yesterday I got some replication trouble.<br>> <br>> My Scenario<br>> <br>> server A < ----- server B <-----> server C <br>> -----> server D<br>
> (dedicated Consumer) (multiple Master replica ID:1) (multiple <br>> Master replica ID:2) (Dedicated Consumer)<br>> <br>> The arrows are depicting the replication directions.<br>> In that scenario everything is fine. <br>
> <br>> But I want to promote server D to a multiple Master replicating from/to <br>> server C.<br>> <br>> On server D I enabled changelog and changed the Replica Role of userRoot <br>> to Multiple Master (now with replica id 3)<br>
> I created a replication aggrement to server C.<br>> <br>> When enabling that I see these error messages in error log on server D. At <br>> 08:51:53 I enabled the replication agreement from server D to server C:<br>
> <br>> [13/Jul/2011:08:49:41 +0200] - 389-Directory/1.2.5 B2010.120.1414 starting <br>> up<br>> [13/Jul/2011:08:49:41 +0200] - slapd started. Listening on All Interfaces <br>> port 389 for LDAP requests<br>
> [13/Jul/2011:08:49:41 +0200] - Listening on All Interfaces port 636 for <br>> LDAPS requests<br>> [13/Jul/2011:08:51:53 +0200] NSMMReplicationPlugin - <br>> repl_set_mtn_referrals: could not set referrals for replica <br>
> dc=mydomain,dc=com: 32<br>> [13/Jul/2011:08:51:53 +0200] NSMMReplicationPlugin - <br>> multimaster_be_state_change: replica dc=mydomain,dc=com is going offline; <br>> disabling replication<br>> [13/Jul/2011:08:51:53 +0200] - WARNING: Import is running with <br>
> nsslapd-db-private-import-mem on; No other process is allowed to access <br>> the database<br>> [13/Jul/2011:08:51:56 +0200] - import userRoot: Workers finished; cleaning <br>> up...<br>> [13/Jul/2011:08:51:57 +0200] - import userRoot: Workers cleaned up.<br>
> [13/Jul/2011:08:51:57 +0200] - import userRoot: Indexing complete. <br>> Post-processing...<br>> [13/Jul/2011:08:51:57 +0200] - import userRoot: Flushing caches...<br>> [13/Jul/2011:08:51:57 +0200] - import userRoot: Closing files...<br>
> [13/Jul/2011:08:51:59 +0200] - import userRoot: Import complete. Processed <br>> 772 entries in 5 seconds. (154.40 entries/sec)<br>> [13/Jul/2011:08:51:59 +0200] NSMMReplicationPlugin - <br>> multimaster_be_state_change: replica dc=mydomain,dc=com is coming online; <br>
> enabling replication<br>> [13/Jul/2011:09:11:00 +0200] NSMMReplicationPlugin - <br>> agmt="cn=server-d_to_server-c" (server-c:389): Unable to aquire replica: <br>> the replica has the same Replica ID as this one. Replication is aborting.<br>
> [13/Jul/2011:09:11:00 +0200] NSMMReplicationPlugin - <br>> agmt="cn=server-d_to_server-c" (server-c:389): Incremental update failed <br>> and requires administrator action<br>> <br>> It says that it has the same replica id, but this is not true. I assigned <br>
> replica ID 3 to server-d and replica ID 2 to server-c<br>> <br>> from dse.ldif of server-d:<br>> dn: cn=replica,cn="dc=mydomain, dc=com",cn=mapping tree,cn=config<br>> objectClass: nsDS5Replica<br>
> objectClass: top<br>> nsDS5ReplicaRoot: dc=mydomain, dc=com<br>> nsDS5ReplicaType: 3<br>> nsDS5Flags: 1<br>> nsds5ReplicaPurgeDelay: 604800<br>> nsDS5ReplicaBindDN: cn=replication manager,cn=config<br>
> cn: replica<br>> creatorsName: <br>> uid=admin,ou=administrators,ou=topologymanagement,o=netscaperoot<br>> modifiersName: <br>> uid=admin,ou=administrators,ou=topologymanagement,o=netscaperoo<br>> t<br>
> createTimestamp: 20110713065052Z<br>> modifyTimestamp: 20110713071007Z<br>> nsDS5ReplicaId: 3<br>> nsState:: //8AAExAHU4AAAAAAAAAAAEAAAA=<br>> nsDS5ReplicaName: 6c5f7682-1dd211b2-abcfd815-dc8d0000<br>> numSubordinates: 1<br>
> <br>> <br>> from dse.ldif if server-c:<br>> dn: cn=replica,cn="dc=mydomain, dc=com",cn=mapping tree,cn=config<br>> objectClass: nsDS5Replica<br>> objectClass: top<br>> nsDS5ReplicaRoot: dc=mydomain, dc=com<br>
> nsDS5ReplicaType: 3<br>> nsDS5Flags: 1<br>> nsDS5ReplicaId: 2<br>> nsds5ReplicaPurgeDelay: 604800<br>> nsDS5ReplicaBindDN: cn=replication manager,cn=config<br>> cn: replica<br>> creatorsName: <br>> uid=admin,ou=administrators,ou=topologymanagement,o=netscaperoot<br>
> modifiersName: cn=Multimaster Replication Plugin,cn=plugins,cn=config<br>> createTimestamp: 20101105122639Z<br>> modifyTimestamp: 20110713073144Z<br>> nsState:: AgAAAMpJHU4AAAAAFgAAAAEAAAA=<br>> nsDS5ReplicaName: ee6ae602-1dd111b2-ae7bc446-6a0e0000<br>
> numSubordinates: 2<br>> <br>> I already erased and recreated the server-d for a couple of times. Always <br>> with the same error.<br>> I created an empty instance of server-d set it to be a dedicated consumer. <br>
> Created replication<br>> aggrement from server-c to server-d. Initiliazed consumer (server-d). On <br>> server-d I enabled<br>> than changelog and changed replication role to multiple master and <br>> assigned replica id 3.<br>
> I when I create the replication aggreement from server-d to server-c I get <br>> this error.<br>> <br>> What am I doing wrong? I did the same thing a while ago for server-b and <br>> server-c and there<br>
> it worked without trouble. (server-b was initially the single master in my <br>> net before<br>> the other servers appeared).<br>> <br>> Thanks for your help,<br>> <br>> Roland<br></div>