Creating Replica 10.3.1 endless

4137
10
Jump to solution
02-10-2016 08:18 AM
RichardHughes2
Occasional Contributor III

Hello,

I've been using Desktop 10.3.1 and Server 10.3.1 to replicate an sde geodatabase.  Here are the steps I took.  Add GUIDs, enable editor tracking, register as versioned, export to xml --> copy to azure vm, load into sde, register as versioned, publish geodata service, add datasets to new mxd on VM, publish map as map service to the same location as geodata service and with the same name, ---> On local machine add map service to mxd, use the distributed geodatabase toolbar to create a 2-way replica to the local sde geodatabase.

The creating replica window has been active forever without any sign of ending. 

Any ideas?

I've been experimenting with different methods and so far I have not been able to register existing data as a replica. 

0 Kudos
1 Solution

Accepted Solutions
JakeSkinner
Esri Esteemed Contributor

I usually recommend to create a few replicas rather than just one large one.  One reason is that if the replica were to become corrupt, you are stuck and unable to sync any of your data.  Another reason, is that you may need to sync changes for some feature classes, but not others.

I would recommend creating a replica for a group of feature classes that are relevant to a project.  For example, if 5 feature classes participate in a web application, you may want to create a single replica for these feature classes.  When updates are ready for the application, you can go ahead an sync the changes.

View solution in original post

10 Replies
JakeSkinner
Esri Esteemed Contributor

Hi Richard,

I would recommend modifying your steps slightly.  After you publish the geodata service, try the following:

1.  On the local machine, add all the feature classes you would like to participate in the replica to ArcMap

2.  Click 'Create Replica' on the Distributed Geodatabase toolbar

3.  Select Two-way, and specify the option to 'Register Existing Data Only'

4.  Select the GeoData service as the geodatabase to replicate to

You should now have a Two-way replica created between the two geodatabases.  You can synchronize changes both ways from your local machine.  Is this what you are looking for?

0 Kudos
RichardHughes2
Occasional Contributor III

Hi Jake,

I just did those steps but on the 'Advanced Create Replica Options' window there are no feature classes listed for me to check-out.  Since this is a two-way replica and not a check-out replica does this matter?  It seems odd to have an empty table there.

Thanks

0 Kudos
JakeSkinner
Esri Esteemed Contributor

This screen should be empty since you chose the option 'Register Existing Data Only'.  You won't be exporting any data.  This option searches the other geodatabase for identical feature classes and then builds the replica.

0 Kudos
RichardHughes2
Occasional Contributor III

Jake,

<ReplicaMsg time='2/10/2016 12:31:52 PM' type='LOG_MESSAGE_TYPE_INFO2' code='90014' elapsed='0.003000 seconds' method='GeoDBReplication::CreateChildReplicaEx' objectClassName='' replicaName='phl_azure'>Creating xml child replica.</ReplicaMsg>

This is the last log entry and the 'Creating Replica' window is still open and I have to end the ArcMap process.

Thanks,

0 Kudos
JakeSkinner
Esri Esteemed Contributor

How many feature classes will be participating in the replica?

0 Kudos
RichardHughes2
Occasional Contributor III

There could be around 100, but a lot are empty.

0 Kudos
RichardHughes2
Occasional Contributor III

I ended the ArcMap Process then reopened the app.  Right clicking on the sde connection --> distributed geodatabase --> synchronize changes works. 

It worrries me that I had to terminate ArcMap before it told me that it had completed the replication process. 

Would you recommend that I create a geodata service for each dataset in the geodatabase and have a replica for each one?

0 Kudos
RichardHughes2
Occasional Contributor III

Hi Jake,

The large number of feature classes looks to be the culprit.  Could you verify whether this workflow is acceptable to handle such a large geodatabase?

1.  Create geodata service from azure vm

2.  on local machine add one dataset to the map and create a 2-way replica named after the dataset

3.  repeat the process for each dataset

0 Kudos
JakeSkinner
Esri Esteemed Contributor

I usually recommend to create a few replicas rather than just one large one.  One reason is that if the replica were to become corrupt, you are stuck and unable to sync any of your data.  Another reason, is that you may need to sync changes for some feature classes, but not others.

I would recommend creating a replica for a group of feature classes that are relevant to a project.  For example, if 5 feature classes participate in a web application, you may want to create a single replica for these feature classes.  When updates are ready for the application, you can go ahead an sync the changes.