Re: Replication issue after improper shutdown

Rich Megginson <rmeggins@xxxxxxxxxx> · Mon, 02 Dec 2013 11:51:41 -0700



    On 12/02/2013 05:49 AM, Sugantha J
      wrote:

    
          Hi Everyone
           
          I am running 389 DS 1.2.8.2 in CentOS 4.8. I have a multi
            master setup, with 12 LDAP servers. Everything was working
            fine, till one of the boxes (ldapw02) suddenly crashed. When
            it came back up, I see the following in the error log,
           
          [25/Nov/2013:20:26:00 -0500] - 389-Directory/1.2.8.2
            B2013.028.104 starting up
          [25/Nov/2013:20:26:01 -0500] - Detected Disorderly
            Shutdown last time Directory Server was running, recovering
            database.
          [25/Nov/2013:20:26:03 -0500] NSMMReplicationPlugin -
            replica_check_for_data_reload: Warning: data for replica
            o=EmpData does not match the data in the changelog (replica
            data (5293f8a1000000040000) > changelog
            (5293f89b000000080000)). Recreating the changelog
            file. This could affect replication with replica's consumers
            in which case the consumers should be reinitialized.
          [25/Nov/2013:20:26:03 -0500] - slapd started.  Listening
            on All Interfaces port 389 for LDAP requests
          [25/Nov/2013:20:30:32 -0500] NSMMReplicationPlugin -
            changelog program - agmt="cn=ldapw022toroon63dsaw03"
            (toroon63dsaw03:389): CSN 5293f761000000020000 not found, we
            aren't as up to date, or we purged
          [25/Nov/2013:20:30:32 -0500] NSMMReplicationPlugin -
            agmt="cn=ldapw022toroon63dsaw03" (toroon63dsaw03:389): Data
            required to update replica has been purged. The replica must
            be reinitialized.
          [25/Nov/2013:20:30:32 -0500] NSMMReplicationPlugin -
            changelog program - agmt="cn=ldapw022toroon63ldapw03"
            (toroon63ldapw03:389): CSN 5293f761000000020000 not found,
            we aren't as up to date, or we purged
          [25/Nov/2013:20:30:32 -0500] NSMMReplicationPlugin -
            agmt="cn=ldapw022toroon63ldapw03" (toroon63ldapw03:389):
            Data required to update replica has been purged. The replica
            must be reinitialized.
          [25/Nov/2013:20:30:32 -0500] NSMMReplicationPlugin -
            agmt="cn=ldapw022toroon63ldapw03" (toroon63ldapw03:389):
            Incremental update failed and requires administrator action
          [25/Nov/2013:20:30:32 -0500] NSMMReplicationPlugin -
            agmt="cn=ldapw022toroon63dsaw03" (toroon63dsaw03:389):
            Incremental update failed and requires administrator action
          [25/Nov/2013:20:35:30 -0500] NSMMReplicationPlugin -
            changelog program - agmt="cn=ldapw022ldapw01" (ldapw01:389):
            CSN 5293f7f3000000050000 not found, we aren't as up to date,
            or we purged
          [25/Nov/2013:20:35:30 -0500] NSMMReplicationPlugin -
            agmt="cn=ldapw022ldapw01" (ldapw01:389): Data required to
            update replica has been purged. The replica must be
            reinitialized.
          [25/Nov/2013:20:35:30 -0500] NSMMReplicationPlugin -
            agmt="cn=ldapw022ldapw01" (ldapw01:389): Incremental update
            failed and requires administrator action.
           
          Replication seems to work to ‘ldapw02’ from all other
            boxes, but replication from ‘ldapw02’ does not work. The
            issue disappears when I initialize all the neighbors from
            the affected box, but I will have to find the root cause for
            this, since this seems
            to happen very frequently.
        
    
    Your boxes are crashing frequently?

          
          Also, I am not able to diagnose the reason for the crash,
            since installing ‘debuginfo’ package is out of my scope.
        
    
        Why?

          
          I see a similar issue being discussed here,
           
          http://thr3ads.net/fedora-directory-users/2007/10/176314-Re-Cant-locate-CSN-in-Multi-Master-replica
           
          and in response to this discussion, the following bug was
            filed,
           
          https://bugzilla.redhat.com/show_bug.cgi?id=388021
           
          It is mentioned that this is fixed in
            “fedora-ds-base-1.2.0”, so I hope I should have this fix in
            my 389 server 1.2.8.2. Do you have any idea as to  why I am
            still getting this problem?
        
    No.

    
          Also, we did a recent LDAP upgrade from 389 DS 1.1.2 to
            1.2.8.2, after
            which I see this problem happening in one or the other LDAP
            server. Is this related? Any help is appreciated.
        
    
    It's going to be extremely difficult to support 1.2.8.  The oldest supported version (meaning someone
              on the dev team can actually try to install and reproduce
              the problem) is 1.2.11.

              
          Regards
          Sugantha J
           
           
      --
389 users mailing list
389-users@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/389-users
    
    
--
389 users mailing list
389-users@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/389-users