Do the same thing again, except for one extra step.
1) Replicate the data from the dup replica
2) Delete the Dup Replica
3) try to replicate to the Dup Replica (say yes to remove
from replica set)
4) Do three again until each duplicate is gone
5) Create new non-dup replica.
6) Done.
I did pretty much exactly what you described, and it
worked (Yeah!). But there is one thing that concerns me:
In the system table "MSysReplicas" of the new replica in
place, the old dup replicas are properly marked as
removed. However, upon several rounds of synchronization
with other replicas within the set, these replicas didn't
take over this information in their system tables. Is this
information only locally stored? If not, what could be
wrong and do I have to fear negative implications?
>-----Originalnachricht-----
>Do the same thing again, except for one extra step.
[quoted text clipped - 27 lines]
>>
>.
Cheval - 21 Sep 2004 23:04 GMT
Well I've also wondered in the past what negative impact
these "Phantom" replicas could do, but as each replica has
a unique ID, the only problem I can see is viewing deleted
replicas in the replica set for the poor human might get
confusing. ie. It's more of a bother for you than the
replication system as such.
Unless of course I'm wrong and please correct me if so.
>-----Original Message-----
>I did pretty much exactly what you described, and it
[quoted text clipped - 42 lines]
>>
>.