Establishing Bi-Directional Replication in Centera

Hi Waheed,

   

                 The steps required to set up bidirectional replication between two clusters. Essentially one cluster is configured to replicate to a second and then the second is configured to replicate to the first.

The examples show replication using application pool. This is the solution recommended by EMC.

Creating the pool manually will not result in successful replication, as the Pool ID needs to be the same across the Clusters.When using profile name and secret to authenticate, the password must be the same on all clusters in the chain.  When using PEA files for authentication, this is not mandatory but you do need to generate and download a PEA file.

You can change the profile secret and download the PEA file using the update profile command.

With GE and CE+ systems, EMC recommends that the replicating clusters have the same compliance mode and settings.

Bidirectional and unidirectional replication is supported in a Centera Virtual Archive (CVA) configuration.

Single Application Using a Pool in a Bidirectional Replication Configuration

1. Choose a cluster to use as the source cluster. The other automatically becomes the target or replication cluster. Work on this first.

On the target or replication cluster:Create a pool for the application to use, unless a pool already exists.

create pool

2.Create an access profile for the application with secret. 

create profile

3.Grant the application profile the minimum set of capabilities for the pool to function on the target cluster either by assigning the pool created in step 1 as its home pool and/or granting it specific capabilities.

See set grants if you choose to grant specific capabilities.

Optionally generate and download a PEA file for the application profile to the local workstation.

4.Export the pool and profile definitions and download it to the local workstation. 

export poolprofilesetup

To facilitate the configuration and minimize the chance for errors, when the Export based on pool or profile selection? prompt appears, choose to export the associated application profile.

5. Where required, send the connection addresses for the cluster, and authentication information or PEA file for the application profile to the system administrator of the first Cluster in the chain.

On the source cluster:

1.Using the export file created in step 3, import the pool and profile configuration as defined on the final cluster in the chain.

import poolprofilesetup

2.Optional - if not done in Step 2, create an Access Profile for the application with secret.  Grant the application profile the minimum set of capabilities for the pool to function on the target cluster either by assigning it as its home pool and/or granting it specific capabilities.

If you have chosen not to use PEA files for authentication, you should synchronize the secrets on all Clusters in the chain.

Optionally generate and download a PEA file for the application profile to the local workstation (optional).

1. Merge all pea files associated with the application for the entire cluster chain.
2. The target cluster now becomes the source cluster and vice-versa.
3. Repeat steps 2 to 7.
4. Where required, send the connection addresses for the clusters and authentication information or PEA file for the application profile to the system administrator for the SDK.
5. Configure replication using the set cluster replication command.

set cluster replication

1. Verify that replication is working properly:

show replication detail command

Hope this helps.

Regards

Satish.N.Kutty