directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Amod Kadam (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (DIRSERVER-1704) Inconistency in Master-Slave Replication
Date Mon, 28 May 2012 15:05:22 GMT

    [ https://issues.apache.org/jira/browse/DIRSERVER-1704?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13284425#comment-13284425
] 

Amod Kadam commented on DIRSERVER-1704:
---------------------------------------

Thanks for your response. 

1) I have created custom partition in MASTER (Virtual Machine)  and SLAVE both running on
two different machines on a network.
2) Set the replication for custom partition in SLAVE but data in the custom partition does
not get replicated. Irrespective of what you set in  'ads-searchbasedn' on the SLAVE  ApacheDS
always replicates data from ou=users,ou=system. 

Here is the configuration on the SLAVE side.

version: 1

dn: ads-replconsumerid=10,ou=replConsumers,ads-serverId=ldapServer,ou=servers,ads-directoryServiceId=default,ou=config
objectclass: ads-replConsumer
objectclass: ads-base
objectclass: top
ads-replaliasderefmode: never
ads-replattributes: *
ads-replconsumerid: 10
ads-replprovhostname: 192.168.2.23
ads-replprovport: 10389
ads-replrefreshinterval: 6000
ads-replrefreshnpersist: TRUE
ads-replsearchfilter: (objectClass=*)
ads-replsearchscope: sub
ads-replsearchsizelimit: 0
ads-replsearchtimeout: 0
ads-repluserdn: uid=admin,ou=system
ads-repluserpassword:: c2VjcmV0
ads-searchbasedn: dc=mypartition,dc=com
ads-enabled: TRUE
ads-replstrictcertvalidation: FALSE
ads-replusetls: FALSE

3) Here is the configuration for the Custom Partition

version: 1

dn: dc=mypartition,dc=com
objectClass: extensibleObject
objectClass: organizationalUnit
objectClass: top
ou: mypartition
dc: mypartition

dn: cn=amod,dc=mypartition,dc=com
objectClass: person
objectClass: top
cn: amod
sn: kadam_update_2




Is there something missing from configuration part ?

4) I can see the branch (ou=consumers,ou=system) on both MASTER as well as SLAVE . How does
it work ? 

The following are the entries on MASTER first followed by SLAVE. Why does both have 'ads-searchbasedn:
ou=system' ? Why both entries are same ?

Entry on MASTER

=================================================
version: 1

dn: ads-dsReplicaId=1,ou=consumers,ou=system
objectclass: ads-base
objectclass: ads-replEventLog
objectclass: top
ads-dsreplicaid: 1
ads-replaliasderefmode: never
ads-replrefreshnpersist: true
ads-replsearchfilter: (objectClass=*)
ads-replsearchscope: sub
ads-searchbasedn: ou=system


=================================================

Entry on SLAVE
=================================================
version: 1

dn: ads-dsReplicaId=1,ou=consumers,ou=system
objectclass: ads-base
objectclass: ads-replEventLog
objectclass: top
ads-dsreplicaid: 1
ads-replaliasderefmode: never
ads-replrefreshnpersist: true
ads-replsearchfilter: (objectClass=*)
ads-replsearchscope: sub
ads-searchbasedn: ou=system


=================================================







                
> Inconistency in Master-Slave Replication
> ----------------------------------------
>
>                 Key: DIRSERVER-1704
>                 URL: https://issues.apache.org/jira/browse/DIRSERVER-1704
>             Project: Directory ApacheDS
>          Issue Type: Bug
>    Affects Versions: 2.0.0-M6
>         Environment: OS : Windows XP Professional Version 2002, Service Pack 3
> Hardware : Intel 1.73 Ghz, 2 GB RAM
>            Reporter: Amod Kadam
>            Priority: Critical
>              Labels: replication
>
> We are testing master-slave configuration for ApacheDS and observed inconsistency when
data from MASTER gets 
> deleted. The record gets deleted from MASTER but does not get deleted from SLAVE. However
this does not happen 
> always. Secondly we are unable to find out the pattern for the same.
> We have seen this issue after executing the Test Cases in sequence. 
> Scenario - Master changes state from RUNNING to STOP and then RUNNING again while Slave
is in 'RUNNING' state
> Ensure that MASTER and SLAVE are running before executing each test case.
> ========================================================
> Test Case No: 1
> 1) Stop Master
> 2) Restart MASTER
> 3) Add an entry to the MASTER 
> dn: cn=amodkadam,ou=users,ou=system
> objectClass: person
> objectClass: top
> sn: kadam
> cn: amodkadam	
> Expected Result : 
> SLAVE should get  corresponding entry for 
> dn: cn=amodkadam,ou=users,ou=system	
> Actual Test Result:
> Same as expected result.
> ===========================================
> Test Case No:2
> 1) Stop Master
> 2) Restart MASTER
> 3) UPDATE an entry in the MASTER 
> dn: cn=amodkadam,ou=users,ou=system
> objectClass: person
> objectClass: top
> sn: kadam_update
> cn: amodkadam	
> Expected Result : 
> SLAVE should have  UPDATED entry for 
> dn: cn=amodkadam,ou=users,ou=system	
> Actual Test Result:
> Same as expected
> ================================================
> 1) Stop Master
> 2) Restart MASTER
> 3) DELETE an entry from the MASTER 
> dn: cn=amodkadam,ou=users,ou=system
> objectClass: person
> objectClass: top
> sn: KADAM
> cn: amodkadam	
> Expected Result : 
> Corresponding entry should be deleted from SLAVE but does not get DELETED
> Actual Result 
> Fails and the record in the SLAVE does not get deleted.
> However this does not happen alwyas. 
> =========================================================================
> Could you let us know what could be issue and how to fix it ?

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message