[Freeipa-users] Attempting to re-provision previous replica

Rich Megginson rmeggins at redhat.com
Thu Oct 23 13:34:56 UTC 2014 

On 10/23/2014 07:01 AM, John Desantis wrote:
> Rob and Rich,
>
>>> ipa-replica-manage del should have cleaned things up. You can clear out
>>> old RUVs with ipa-replica-manage too via list-ruv and clean-ruv. You use
>>> list-ruv to get the id# to clean and clean-ruv to do the actual cleaning.
>> I remember having previously tried this task, but it had failed on
>> older RUV's which were not even active (the KDC was under some strain
>> so ipa queries were timing out).  However, I ran it again and have
>> been able to delete all but 1 (it's still running) RUV referencing the
>> previous replica.
>>
>> I'll report back once the tasks finishes or fails.
> The last RUV is "stuck" on another replica.  It fails with the following error:
>
> [23/Oct/2014:08:55:09 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Initiating CleanAllRUV Task...
> [23/Oct/2014:08:55:10 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Retrieving maxcsn...
> [23/Oct/2014:08:55:10 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Found maxcsn (5447f861000000180000)
> [23/Oct/2014:08:55:10 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Cleaning rid (24)...
> [23/Oct/2014:08:55:10 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Waiting to process all the updates from the deleted replica...
> [23/Oct/2014:08:55:10 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Waiting for all the replicas to be online...
> [23/Oct/2014:08:55:10 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Waiting for all the replicas to receive all the deleted replica
> updates...
> [23/Oct/2014:08:55:11 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Replica maxcsn (5447f56b000200180000) is not caught up with deleted
> replica's maxcsn(5447f861000000180000)
> [23/Oct/2014:08:55:11 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Replica not caught up (agmt="cn=meToiparepbackup.our.personal.domain"
> (iparepbackup:389))
> [23/Oct/2014:08:55:11 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Not all replicas caught up, retrying in 10 seconds
> [23/Oct/2014:08:55:23 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Replica maxcsn (5447f56b000200180000) is not caught up with deleted
> replica's maxcsn(5447f861000000180000)
> [23/Oct/2014:08:55:23 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Replica not caught up (agmt="cn=meToiparepbackup.our.personal.domain"
> (iparepbackup:389))
> [23/Oct/2014:08:55:23 -0400] NSMMReplicationPlugin - CleanAllRUV Task:
> Not all replicas caught up, retrying in 20 seconds
>
> I then abort the task since the retrying went up to 14400 seconds.

Mark, do you know what is going on here?

>
> Would this be a simple re-initialization from the master on the host
> "iparepbackup"?
>
> Thanks,
> John DeSantis
>
> 2014-10-22 16:03 GMT-04:00 John Desantis <desantis at mail.usf.edu>:
>> Rob and Rich,
>>
>>> ipa-replica-manage del should have cleaned things up. You can clear out
>>> old RUVs with ipa-replica-manage too via list-ruv and clean-ruv. You use
>>> list-ruv to get the id# to clean and clean-ruv to do the actual cleaning.
>> I remember having previously tried this task, but it had failed on
>> older RUV's which were not even active (the KDC was under some strain
>> so ipa queries were timing out).  However, I ran it again and have
>> been able to delete all but 1 (it's still running) RUV referencing the
>> previous replica.
>>
>> I'll report back once the tasks finishes or fails.
>>
>> Thanks,
>> John DeSantis
>>
>>
>> 2014-10-22 15:49 GMT-04:00 Rob Crittenden <rcritten at redhat.com>:
>>> Rich Megginson wrote:
>>>> On 10/22/2014 12:55 PM, John Desantis wrote:
>>>>> Richard,
>>>>>
>>>>>> You should remove the unused ruv elements.  I'm not sure why they
>>>>>> were not
>>>>>> cleaned.  You may have to use cleanallruv manually.
>>>>>> https://access.redhat.com/documentation/en-US/Red_Hat_Directory_Server/9.0/html/Administration_Guide/Managing_Replication-Solving_Common_Replication_Conflicts.html#cleanruv
>>>>>>
>>>>>>
>>>>>> note - use the cleanallruv procedure, not cleanruv.
>>>>> I'll try that, thanks for the guidance.
>>>>>
>>>>>> What is the real problem you have?  Did replication stop working? Are
>>>>>> you
>>>>>> getting error messages?
>>>>> I cannot get the host to be a replica.  Each time I run
>>>>> `ipa-replica-install
>>>>> replica-info-host-in-question.our.personal.domain.gpg' it fails.  I
>>>>> had assumed it was due to the fact that the host was already a
>>>>> replica, but had to be taken offline due to a hard disk failing.  The
>>>>> machine was re-provisioned after the new hard drive was installed.
>>>> Ok.  I don't know if we have a documented procedure for that case. I
>>>> assumed that if you first ran ipa-replica-manage del, then
>>>> ipa-replica-prepare, then ipa-replica-install, that would take care of
>>>> that.
>>> ipa-replica-manage del should have cleaned things up. You can clear out
>>> old RUVs with ipa-replica-manage too via list-ruv and clean-ruv. You use
>>> list-ruv to get the id# to clean and clean-ruv to do the actual cleaning.
>>>
>>>>> When I enabled extra debugging during the installation process, the
>>>>> initial error was that the dirsrv instance couldn't be started.  I
>>>>> checked into this and found that there were missing files in
>>>>> /etc/dirsrv/slapd-BLAH directory.  I was then able to start dirsrv
>>>>> after copying some schema files from another replica.  The install did
>>>>> move forward but then failed with Apache and its IPA configuration.
>>>>>
>>>>> I performed several uninstalls and re-installs, and at one point I got
>>>>> error code 3 from ipa-replica-install, which is why I was thinking
>>>>> that the old RUV's and tombstones were to blame.
>>>> It could be.  I'm really not sure what the problem is at this point.
>>> I think we'd need to see ipareplica-install.log to know for sure. It
>>> could be the sort of thing where it fails early but doesn't kill the
>>> install, so the last error is a red herring.
>>>
>>> rob
>>>
>>>>> Thanks,
>>>>> John DeSantis
>>>>>
>>>>>
>>>>> 2014-10-22 12:51 GMT-04:00 Rich Megginson <rmeggins at redhat.com>:
>>>>>> On 10/22/2014 10:31 AM, John Desantis wrote:
>>>>>>> Richard,
>>>>>>>
>>>>>>> You helped me before in #freeipa, so I appreciate the assistance again.
>>>>>>>
>>>>>>>> What version of 389 are you using?
>>>>>>>> rpm -q 389-ds-base
>>>>>>> 389-ds-base-1.2.11.15-34.el6_5
>>>>>>>
>>>>>>> Thanks,
>>>>>>> John DeSantis
>>>>>>>
>>>>>>> 2014-10-22 12:09 GMT-04:00 Rich Megginson <rmeggins at redhat.com>:
>>>>>>>> On 10/22/2014 09:42 AM, John Desantis wrote:
>>>>>>>>> Hello all,
>>>>>>>>>
>>>>>>>>> First and foremost, a big "thank you!" to the FreeIPA developers
>>>>>>>>> for a
>>>>>>>>> great product!
>>>>>>>>>
>>>>>>>>> Now, to the point!
>>>>>>>>>
>>>>>>>>> We're trying to re-provision a previous replica using the standard
>>>>>>>>> documentation via the Red Hat site:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Identity_Management_Guide/Setting_up_IPA_Replicas.html
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> However, we're running into errors during the import process.  The
>>>>>>>>> errors are varied and fail at random steps; there was an issue with
>>>>>>>>> NTP or HTTP or LDAP, etc.  This did not happen when we promoted a
>>>>>>>>> separate node to become a replica.
>>>>>>>>>
>>>>>>>>> We had previously removed the replica via `ipa-replica-manage del`
>>>>>>>>> and
>>>>>>>>> ensured that no trace of it being a replica existed: removed DNS
>>>>>>>>> records and verified that the host enrollment was not present.  I did
>>>>>>>>> not use the "--force" and "--cleanup" options.
>>>>>>>> What version of 389 are you using?
>>>>>>>> rpm -q 389-ds-base
>>>>>> You should remove the unused ruv elements.  I'm not sure why they
>>>>>> were not
>>>>>> cleaned.  You may have to use cleanallruv manually.
>>>>>> https://access.redhat.com/documentation/en-US/Red_Hat_Directory_Server/9.0/html/Administration_Guide/Managing_Replication-Solving_Common_Replication_Conflicts.html#cleanruv
>>>>>>
>>>>>>
>>>>>> note - use the cleanallruv procedure, not cleanruv.
>>>>>>
>>>>>>>>> When I check RUV's against the host in question, there are
>>>>>>>>> several.  I
>>>>>>>>> also queried the tombstones against the host and found two entries
>>>>>>>>> which have valid hex time stamps;  coincidentally, out of the 9
>>>>>>>>> tombstone entries, 2 have "nsds50ruv" time stamps.  I'll paste
>>>>>>>>> sanitized output below:
>>>>>>>>>
>>>>>>>>> # ldapsaerch -x -W -LLL -D "cn=directory manager" -b
>>>>>>>>> "dc=our,dc=personal,dc=domain" '(objectclass=nsTombstone)'
>>>>>>>>> Enter LDAP Password:
>>>>>>>>> dn:
>>>>>>>>>
>>>>>>>>> nsuniqueid=ffffffff-ffffffff-ffffffff-ffffffff,dc=our,dc=personal,dc=domain
>>>>>>>>>
>>>>>>>>> objectClass: top
>>>>>>>>> objectClass: nsTombstone
>>>>>>>>> objectClass: extensibleobject
>>>>>>>>> nsds50ruv: {replicageneration} 50ef13ae000000040000
>>>>>>>>> nsds50ruv: {replica 4 ldap://master.our.personal.domain:389}
>>>>>>>>> 5164d147000000040000 5447bda 8000100040000
>>>>>>>>> nsds50ruv: {replica 22 ldap://separatenode.our.personal.domain:389}
>>>>>>>>> 54107f9f000000160000 54436b 25000000160000
>>>>>>>>> nsds50ruv: {replica 21 ldap://iparepbackup.our.personal.domain:389}
>>>>>>>>> 51b734de000000150000 51b7 34ef000200150000
>>>>>>>>> nsds50ruv: {replica 19
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 510d56c9000100130000
>>>>>>>>> 510d82 be000200130000
>>>>>>>>> nsds50ruv: {replica 18
>>>>>>>>> ldap://host-in-question.our.personal.domain:389}
>>>>>>>>> nsds50ruv: {replica 17
>>>>>>>>> ldap://host-in-question.our.personal.domain:389}
>>>>>>>>> nsds50ruv: {replica 16
>>>>>>>>> ldap://host-in-question.our.personal.domain:389}
>>>>>>>>> nsds50ruv: {replica 15
>>>>>>>>> ldap://host-in-question.our.personal.domain:389}
>>>>>>>>> nsds50ruv: {replica 14
>>>>>>>>> ldap://host-in-question.our.personal.domain:389}
>>>>>>>>> nsds50ruv: {replica 13
>>>>>>>>> ldap://host-in-question.our.personal.domain:389}
>>>>>>>>> nsds50ruv: {replica 12
>>>>>>>>> ldap://host-in-question.our.personal.domain:389}
>>>>>>>>> nsds50ruv: {replica 23
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 54187702000200170000
>>>>>>>>> 541878 9a000000170000
>>>>>>>>> dc: our
>>>>>>>>> nsruvReplicaLastModified: {replica 4
>>>>>>>>> ldap://master.our.personal.domain:389} 5447bce8
>>>>>>>>> nsruvReplicaLastModified: {replica 22
>>>>>>>>> ldap://separatenode.our.personal.domain:389} 54436a5e
>>>>>>>>> nsruvReplicaLastModified: {replica 21
>>>>>>>>> ldap://iparepbackup.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 19
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 18
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 17
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 16
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 15
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 14
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 13
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 12
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>> nsruvReplicaLastModified: {replica 23
>>>>>>>>> ldap://host-in-question.our.personal.domain:389} 00000000
>>>>>>>>>
>>>>>>>>> dn:
>>>>>>>>>
>>>>>>>>> nsuniqueid=c08a2803-5b5a11e2-a527ce8b-8fa47d35,cn=host-in-question.our.personal.domain,cn=maste
>>>>>>>>>
>>>>>>>>>      rs,cn=ipa,cn=etc,dc=our,dc=personal,dc=domain
>>>>>>>>> objectClass: top
>>>>>>>>> objectClass: nsContainer
>>>>>>>>> objectClass: nsTombstone
>>>>>>>>> cn: host-in-question.our.personal.domain
>>>>>>>>> nsParentUniqueId: e6fa9418-5b5711e2-a1a9825b-daf5b5b0
>>>>>>>>>
>>>>>>>>> dn:
>>>>>>>>>
>>>>>>>>> nsuniqueid=664c4383-6d6311e2-8db6e946-de27dd8d,cn=host-in-question.our.personal.domain,cn=maste
>>>>>>>>>
>>>>>>>>>      rs,cn=ipa,cn=etc,dc=our,dc=personal,dc=domain
>>>>>>>>> objectClass: top
>>>>>>>>> objectClass: nsContainer
>>>>>>>>> objectClass: nsTombstone
>>>>>>>>> cn: host-in-question.our.personal.domain
>>>>>>>>> nsParentUniqueId: e6fa9418-5b5711e2-a1a9825b-daf5b5b0
>>>>>>>>>
>>>>>>>>> As you can see, the "host-in-question" has many RUV's and of which
>>>>>>>>> two
>>>>>>>>> appear to be "active" and two entries which I believe (pardon my
>>>>>>>>> ignorance) possibly correlate with the "active" entries of the
>>>>>>>>> "host-in-question".
>>>>>>>>>
>>>>>>>>> Do these two tombstone entries need to be deleted with ldapdelete
>>>>>>>>> before we can re-provision "host-in-question" and add it back as a
>>>>>>>>> replica?
>>>>>> No, you cannot delete tombstones manually.  They will be cleaned up
>>>>>> at some
>>>>>> point by the dirsrv tombstone reap thread, and they should not be
>>>>>> interfering with anything.
>>>>>>
>>>>>> What is the real problem you have?  Did replication stop working? Are
>>>>>> you
>>>>>> getting error messages?
>>>>>>
>>>>>>>>> Thank you,
>>>>>>>>> John DeSantis
>>>>>>>>>
>>>>>>>> --
>>>>>>>> Manage your subscription for the Freeipa-users mailing list:
>>>>>>>> https://www.redhat.com/mailman/listinfo/freeipa-users
>>>>>>>> Go To http://freeipa.org for more info on the project
>>>>>> --
>>>>>> Manage your subscription for the Freeipa-users mailing list:
>>>>>> https://www.redhat.com/mailman/listinfo/freeipa-users
>>>>>> Go To http://freeipa.org for more info on the project

More information about the Freeipa-users mailing list