Re: [389-users] duplicate replica id detected - can I remove replicas?

[Rich Megginson]

On 08/10/2011 11:00 PM, Amos Shapira wrote:
>  Some more research after I sent my question (narrower search terms) 
>  found the following bug and fix:
>
>  https://bugzilla.redhat.com/show_bug.cgi?id=624442
>
>  The package I use is centos-ds-base-8.1.0-0.14.el5.centos.2
>
>  Does this mean that this is a bug in the package I have and I should 
>  just use different replication id's?
>  Or can I still somehow clear the tombstones? 
You might be able to use db2ldif -r to
dump the database along with the 
replication meta data, then edit the LDIF file to remove the bogus data, 
then ldif2db to reload your database from the LDIF file.
>  Thanks,
>
>  --Amos
>
>  On 11 August 2011 14:55, <389-users-owner(a)lists.fedoraproject.org 
>  <mailto:[e-mail]>> wrote:
>
>      You are not allowed to post to this mailing list, and your message has
>      been automatically rejected.  If you think that your messages are
>      being rejected in error, contact the mailing list owner at
>      389-users-owner(a)lists.fedoraproject.org
>      <mailto:[e-mail]>.
>
>
>
>      ---------- Forwarded message ----------
>      From: Amos Shapira <amos.shapira(a)gmail.com
>      <mailto:[e-mail]>>
>      To: "General discussion list for the 389 Directory server
>      project." <389-users(a)lists.fedoraproject.org
>      <mailto:[e-mail]>>
>      Date: Thu, 11 Aug 2011 14:55:08 +1000
>      Subject: duplicate replica id detected - can I remove replicas?
>      Hello,
>
>      I've deleted old replication agreements and setup new ones as part
>      of server reshuffling (previous consumers turned into multi-masters).
>
>      I'm following the instructions in
>      http://directory.fedoraproject.org/wiki/Howto:WalkthroughMultimasterSSL
>      which look like a shorter summary of the official documentation.
>
>      When I finally click on the "Inisitalize consumer" I get a pop-up
>      about duplicate replicas existing.
>
>      The replica numbers on the servers are "4" (on the supplier I
>      intend to initialise from) and "3" (on the intended consumer).
>
>      I followed instructions given in this forum earlier to list
>      replicas on both servers and these are the results:
>
>      # ./ldapsearch -x -b
>      'nsuniqueid=ffffffff-ffffffff-ffffffff-ffffffff,dc=company' -D
>      'cn=Directory Manager' -w ******* -h cunsumers-ip-addr -s base
>      objectclass=nstombstone
>      version: 1
>      dn: nsuniqueid=ffffffff-ffffffff-ffffffff-ffffffff,dc=company
>      objectClass: top
>      objectClass: nsTombstone
>      objectClass: extensibleobject
>      nsds50ruv: {replicageneration} 4a679e1c000000010000
>      nsds50ruv: {replica 3 ldap://directory2:389
>       }
>      nsds50ruv: {replica 1 ldap://directory5:389} 4a679ebf000000010000
>       4e30ce4e000200010000
>      nsds50ruv: {replica 65535 ldap://directory2:389}
>      4e323b300000ffff0000 4e4026ef0000ffff0000
>      dc: company
>      nsruvReplicaLastModified: {replica 3 ldap://directory2:389} 00000000
>      nsruvReplicaLastModified: {replica 1 ldap://directory5:389} 000000
>       00
>      nsruvReplicaLastModified: {replica 65535 ldap://directory2:389}
>      4e4026d9
>
>      # ./ldapsearch -x -b
>      'nsuniqueid=ffffffff-ffffffff-ffffffff-ffffffff,dc=company' -D
>      'cn=Directory Manager' -w ******** -s base objectclass=nstombstone
>      version: 1
>      dn: nsuniqueid=ffffffff-ffffffff-ffffffff-ffffffff,dc=company
>      objectClass: top
>      objectClass: nsTombstone
>      objectClass: extensibleobject
>      nsds50ruv: {replicageneration} 4a679e1c000000010000
>      nsds50ruv: {replica 3 ldap://directory2:389}
>      nsds50ruv: {replica 1 ldap://directory5:389} 4a679ebf000000010000
>       4e30ce4e000200010000
>      nsds50ruv: {replica 65535 ldap://directory2:389} 4e4026ef0000ffff0000
>      dc: company
>      nsruvReplicaLastModified: {replica 3 ldap://directory2:389} 00000000
>      nsruvReplicaLastModified: {replica 1 ldap://directory5:389} 4e30ce
>       27
>      nsruvReplicaLastModified: {replica 65535 ldap://directory2:389}
>      4e4026da
>
>      directory1 is the server I'm running the query on and is the
>      intended initialising supplier.
>      directory2 is the server I'm trying to initialize.
>      directory1 and directory2 are the servers I intend to be the new
>      multi-masters. I configured them to use LDAPS on port 636 with
>      certificates.
>      directory5 is one of the two older multi-masters (the old
>      multi-masters haven't been configured corretly, I guess that's why
>      we don't see a second replica from the old multi-master pair).
>
>      Is there a way for me to remove the old replicas or must I just
>      pick new replica numbers for the new multi-masters?
>
>      In case this matters - I'm configuring things through a
>      centos-idm-console which I copied to my Ubuntu desktop in another
>      location and accessing the DC over VPN. I have direct access from
>      my desktop to ports 636 and 389 on both servers. When I setup the
>      replication agreement the console complains that it can't validate
>      the connection but I suspect this is related to it not having
>      access to the cert8.db file
>      (https://bugzilla.redhat.com/show_bug.cgi?id=200989).
>
>      Thanks,
>
>      --Amos
>
>

duplicate replica id detected - can I remove replicas?

[Penedo]

On 11 August 2011 23:57, Rich Megginson <rmeggins(a)redhat.com&gt; wrote:

>  **
>  On 08/10/2011 11:00 PM, Amos Shapira wrote:
>
>  Some more research after I sent my question (narrower search terms) found
>  the following bug and fix:
>
>  https://bugzilla.redhat.com/show_bug.cgi?id=624442
>
>  The package I use is centos-ds-base-8.1.0-0.14.el5.centos.2
>
>  Does this mean that this is a bug in the package I have and I should just
>  use different replication id's?
>  Or can I still somehow clear the tombstones?
>
>  You might be able to use db2ldif -r to dump the database along with the
>  replication meta data, then edit the LDIF file to remove the bogus data,
>  then ldif2db to reload your database from the LDIF file.
>  
Thanks. This sounds like what I was after.

For now I just assigned new replica numbers to each server and it seems to
work.
Is there any reason to risk the data in order to drop the old replica ID's
or is it OK to ignore them and just leave them there?

Cheers,

--Amos

duplicate replica id detected - can I remove replicas?

[Rich Megginson]

On 08/12/2011 01:05 AM, Penedo wrote:
>  On 11 August 2011 23:57, Rich Megginson <rmeggins(a)redhat.com 
>  <mailto:[e-mail]>> wrote:
>
>      On 08/10/2011 11:00 PM, Amos Shapira wrote:
> >     Some more research after I sent my question (narrower search
> >     terms) found the following bug and fix:
> >
> >     https://bugzilla.redhat.com/show_bug.cgi?id=624442
> >
> >     The package I use is centos-ds-base-8.1.0-0.14.el5.centos.2
> >
> >     Does this mean that this is a bug in the package I have and I
> >     should just use different replication id's?
> >     Or can I still somehow clear the tombstones?
>      You might be able to use db2ldif -r to dump the database along
>      with the replication meta data, then edit the LDIF file to remove
>      the bogus data, then ldif2db to reload your database from the LDIF
>      file.
>
>
>  Thanks. This sounds like what I was after.
>
>  For now I just assigned new replica numbers to each server and it 
>  seems to work.
>  Is there any reason to risk the data in order to drop the old replica 
>  ID's or is it OK to ignore them and just leave them there? 
It shouldn't
cause any data loss.  You might find annoying messages in 
the errors log though.
>  Cheers,
>
>  --Amos