[Freeipa-users] understanding RUVs?

[thierry bordaz]

On 04/21/2015 09:11 AM, Martin Kosek wrote:
> On 04/21/2015 01:26 AM, Janelle wrote:
>> Hello,
>>
>> When I was working with OpenLDAP, and AD - and did not deal with "RUV"s the way
>> I am with 389-ds and IPA.
>>
>> I am trying to understand what is "normal" for values. If I am looking at this
>> (and seem to have no replication problems):
>>
>> ipa-replica-manage list-ruv
>>
>> ipa001.example.com:389: 13
>> ipa002.example.com:389: 12
>> ipa003.example.com:389: 11
>> ipa004.example.com:389: 10
>> ipa005.example.com:389: 7
>> ipa006.example.com:389: 6
>> ipa007.example.com:389: 5
>> ipa008.example.com:389: 3
>> ipa009.example.com:389: 16
>> ipa00a.example.com:389: 17
>> ipa00b.example.com:389: 15
>> ipa00c.example.com:389: 14
>> ipa00d.example.com:389: 9
>> ipa00e.example.com:389: 8
>> ipa00f.example.com:389: 4
>>
>> I guess I was wondering, should I be seeing all the same values or should they
>> all be unique based on being "replicated" and the order they were added?
> They should be unique, that's for sure. There is some info on them in Red Hat
> DS docs:
>
> https://access.redhat.com/documentation/en-US/Red_Hat_Directory_Server/9.0/html/Administration_Guide/Managing_Replication-Solving_Common_Replication_Conflicts.html#cleanruv
>
> I am just not sure if they are replicated or per-server. But given they live in
> SUFFIX, I assume they are. The list above looks OK to me, so it should not
> cause the replication problems.

Hello,

Yes this RUV is normal.

The RUV is a special 389-ds entry that is per server. This entry allows 
replication protocol (run by the replica agreements)
to detect what updates are missing and then send the missing ones.

The command list-ruv displays a subset of the attribute values of that 
entry. It displays url and the replicaId.
A normal RUV in a replication topology contains unique replicaId and a 
url must be listed only once.


thanks
thierry
>
> But I am rather CCing Thierry to advise here.
>
>> Or is
>> it telling me something else? Sorry, I guess I am still trying to wrap my head
>> around replication metadata.