On 06/12/2017 07:32 AM, Nick Campion via FreeIPA-users wrote:
>
> Thanks Mark,
>
> So this example is a user password change using kinit, the password
> has been changed on freeipa02 but not then replicated to the others.
> This happens for other records, but I don't have examples of these at
> the moment.
>
> As far as I'm aware, there is no fractal replication set up.
>
IPA uses fractional replication, and it's possible these attributes are
ignored/skipped.  To confirm you can run this search on freeipa02:

ldapsearch -D "cn=directory manager" -W -b cn=config -xLLL
objectclass=nsds5ReplicationAgreement

Then please share these entries so we can see how they are configured. 
Perhaps do this on freeipa01 as well for comparison.
>
> Freeipa01:
>
> # dynamic-kepler, users, accounts, ipa.example.com
> dn: uid=dynamic-kepler,cn=users,cn=accounts,dc=ipa,dc=example,dc=com
> uid: dynamic-kepler
> krbLastPwdChange: 20170608170011Z
> krbPasswordExpiration: 20170608170011Z
>
> Freeipa02:
>
> # dynamic-kepler, users, accounts, ipa.example.com
> dn: uid=dynamic-kepler,cn=users,cn=accounts,dc=ipa,dc=example,dc=com
> uid: dynamic-kepler
> krbLastPwdChange: 20170608170021Z
> krbPasswordExpiration: 20170906170021Z
>
> Freeipa03:
>
> # dynamic-kepler, users, accounts, ipa.example.com
> dn: uid=dynamic-kepler,cn=users,cn=accounts,dc=ipa,dc=example,dc=com
> uid: dynamic-kepler
> krbLastPwdChange: 20170608170011Z
> krbPasswordExpiration: 20170608170011Z
>
> Errors on Freeipa02:
>
> [08/Jun/2017:01:46:50.635529447 +0000] replica_generate_next_csn:
> opcsn=5938ac8b000500030000 <= basecsn=5938ac8b000500040000, adjusted
> opcsn=5938ac8b000600030000
> [08/Jun/2017:12:16:46.497249649 +0000] replica_generate_next_csn:
> opcsn=5939402f000500030000 <= basecsn=5939402f000800040000, adjusted
> opcsn=5939402f000900030000
> [08/Jun/2017:23:38:48.197750001 +0000] replica_generate_next_csn:
> opcsn=5939e009000100030000 <= basecsn=5939e009000f00040000, adjusted
> opcsn=5939e009001000030000
>
> The other nodes have no errors from this data.
>
> Access logs:
>
> Freeipa01:
>
> [08/Jun/2017:01:46:50.635529447 +0000] replica_generate_next_csn:
> opcsn=5938ac8b000500030000 <= basecsn=5938ac8b000500040000, adjusted
> opcsn=5938ac8b000600030000
> [08/Jun/2017:12:16:46.497249649 +0000] replica_generate_next_csn:
> opcsn=5939402f000500030000 <= basecsn=5939402f000800040000, adjusted
> opcsn=5939402f000900030000
> [08/Jun/2017:23:38:48.197750001 +0000] replica_generate_next_csn:
> opcsn=5939e009000100030000 <= basecsn=5939e009000f00040000, adjusted
> opcsn=5939e009001000030000
>
This is from an error log :-)
>
> Freeipa02:
>
> Shows no logs "to" the other 2 nodes.
>
Well it would only show incoming connections, not outgoing. 
>
> Freeipa03:
>
> [08/Jun/2017:17:10:06.343697044 +0000] conn=9237 fd=70 slot=70
> connection from 192.168.0.12 to 192.168.0.13
> [08/Jun/2017:19:54:05.025713675 +0000] conn=9665 fd=70 slot=70
> connection from 192.168.0.12 to 192.168.0.13
>
> Freeipa02 replication logging:
>
> [09/Jun/2017:11:24:58.827281135 +0000] NSMMReplicationPlugin -
> csnplCommitALL: processing data csn 593964af000900030000
>
> Repeats 800 - 900 time per second with a different csn.
>
It looks like its replicating to other replicas, but some updates are
skipped.  This again could be fractional replication "working".  

If you look through freeipa01's access log what operation is this csn
from:   5937cccd000f00030000 ?   Could this be one of the password
updates that is not replicated?  This update is not sent to the other
replicas that's why I'm asking.

Thanks,
Mark
>
> On 08/06/17 15:45, Mark Reynolds wrote:
>>
>>
>> On 06/07/2017 10:58 AM, Nick Campion via FreeIPA-users wrote:
>>>
>>> Hi all,
>>>
>>>  
>>>
>>> We have a 3 master setup that is failing to replicate changes from a
>>> particular node to the other IPA instances. The replication status
>>> says it's all fine, however the record hasn't been changed on the
>>> other servers. We've seen this on user password changes, adding
>>> hosts and services. The only thing we've found that seems to fix
>>> this temporarily is to re-initialize from the master with the
>>> changed record. A force-sync doesn't pick up the changed record.
>>>
>> What is the change you making, what attribute are you updating? 
>> Could it be possible that its being excluded by fractional
>> replication?  Or is it all changes?
>>
>> Any errors in the logs on the nodes(good and bad): 
>> /var/log/dirsrv/slapd-INSTANCE/errors
>>
>> Do you see replication sessions starting between the bad node and
>> good ones?  Are they talking?  Check the access log (
>> /var/log/dirsrv/slapd-INSTANCE/access) on a good node and look for
>> "connection from <BAD NODE IP address>"
>>
>> Next would be to enable replication logging on the bad node and
>> reproduce the problem (then disable repl logging right away), then
>> send us the logs to look at.  See 
>> https://access.redhat.com/documentation/en-us/red_hat_directory_server/10/html/administration_guide/managing_replication-troubleshooting_replication_related_problems
>>
>> Regards,
>> Mark
>>
>>> Not sure what logs would be helpful to diagnose what is happening in
>>> this setup. 
>>>
>>> # ipa-replica-manage -v list `hostname`
>>> freeipa03.mgmt.example.com: replica
>>> last init status: None
>>> last init ended: 1970-01-01 00:00:00+00:00
>>> last update status: Error (0) Replica acquired successfully:
>>> Incremental update succeeded
>>> last update ended: 2017-06-07 14:43:53+00:00
>>> freeipa02.mgmt.example.com: replica
>>> last init status: None
>>> last init ended: 1970-01-01 00:00:00+00:00
>>> last update status: Error (0) Replica acquired successfully:
>>> Incremental update succeeded
>>> last update ended: 2017-06-07 14:43:53+00:00
>>>
>>> # ldapsearch -W -x -D "cn=directory manager" -b
>>> "cn=users,cn=accounts,dc=ipa,dc=example,dc=com"
>>> "nsds5ReplConflict=*" \* nsds5ReplConflict
>>> Enter LDAP Password:
>>> # extended LDIF
>>> #
>>> # LDAPv3
>>> # base <cn=users,cn=accounts,dc=ipa,dc=example,dc=com> with scope
>>> subtree
>>> # filter: nsds5ReplConflict=*
>>> # requesting: * nsds5ReplConflict
>>> #
>>>
>>> # search result
>>> search: 2
>>> result: 0 Success
>>>
>>> # numResponses: 1
>>>
>>> Any help in what else can be checked or what logs would be helpful
>>> would be appreciated.
>>>
>>> Thanks
>>>
>>> Nick
>>>
>>>
>>>
>>> _______________________________________________
>>> FreeIPA-users mailing list -- freeipa-users@lists.fedorahosted.org
>>> To unsubscribe send an email to freeipa-users-le...@lists.fedorahosted.org
>>
>
>
>
> _______________________________________________
> FreeIPA-users mailing list -- freeipa-users@lists.fedorahosted.org
> To unsubscribe send an email to freeipa-users-le...@lists.fedorahosted.org

_______________________________________________
FreeIPA-users mailing list -- freeipa-users@lists.fedorahosted.org
To unsubscribe send an email to freeipa-users-le...@lists.fedorahosted.org

Reply via email to