PowerStore: Manage replication session destination storage resources, policies, replication rules and remote systems post unplanned permanent source system failure when initial sync has not completed.

This article will document how to remove restrictions on the destination resource and protection objects so that the destination system can be cleaned up or reconfigured.

---

1.Cause

After a permanent source system failure, replication is usually failed over from the destination system to enable production access to the replication destination and enable policy management. The Failover operation is restricted if the initial synchronization has not completed, and the following error is returned:

            Failover is not allowed.  Possible reasons are that the replication session has not completed an initial
            synchronization, or the volume group being replicated has no member volumes. (0xE0201005000D)

Failover from the replication destination is restricted until the initial synchronization has completed and the replicated data is on the destination. After an unexpected source failure during an initial synchronization, cleanup of the destination resource and policy changes will be blocked because the replication resource is still a replication destination.

The following operations will be restricted on sessions which have not completed the initial synchronizations:

    

•     Delete the destination resource because it has a replicated policy with a replication rule.
  
  
      
•     Delete the policy and/or replication rule because it is a replicated policy and cannot be deleted by the user and the system won’t remove it because it is in use by one or more destination resources.
  
      
•     Delete the remote system specified in the replication rule because it is currently in use by a replicated rule/policy that is attached to a resource.
  
          The following error will be displayed in the Job:
          Failed to delete remote system H0255 as there are existing replication rules between the local and remote systems. If no such rules are visible, this refers to temporary rules created by the system that will be automatically cleaned up. (0xE0201002002F)
  
      
•     Unassign the policy from the resource in the resource table because this operation is restricted on replication destination resources (volumes and volume groups).
  
          The following error will be displayed in the Job: 
          Cannot modify or delete the protection policy on resource "unplan permanent source system failure" because it is a replication destination. If you need to manually remove the policy because replication failed to start then please use the force option. (0xE02020010020)

  ---

  2.Solution

  Since the source system is permanently down, replication cannot be restarted and the initial synchronization will never complete, leaving the destination without the replicated source data. Failover is restricted in this workflow since there is no replicated data.  However, the user may want to remove the replication destination resources and the remote system.
  
  Remote System alerts will be the first indication that there is a source system failure.
  
  Example alerts:
  0x01700304   All data connections from the remote system H0255 are disconnected.
  0x01700203   The remote system H0255 management connection is lost.
  
      
•     If the source is going to come back up - do nothing and the session will recover and continue with the synchronization.
      
•     If the source is permanently down - follow these steps to remove replication from the destination resource allowing management and cleanup of the destination resource.
  
  Policy:

   

  Note: This document specifically covers replications which have not completed the initial synchronization cycle and Failover is restricted.  Other resources may be replicating from the failed source system and those can be failed over to gain production access to the data and enable unassign of a policy.

   

  
  
  For resources with replication which has not completed the initial synchronization:
  
  Navigate to the resource (volume or volume group)->details->Protection KPI in the GUI.  Select the "UNASSIGN POLICY" button.  You must confirm this recovery operation since it could have unexpected results if the source is restored.
  
  
  This will remove the replicated policy from the destination object and remove the replication session.  This recovery operation is only allowed from within the protection KPI of a resource and is not available from the volume or volume group tables.
  
  Destination Resource (volume or volume group):
  
  The resource on the destination is currently a replication destination which restricts some volume or volume group operations. Once the protection policy has been unassigned, the resource can be deleted or converted to a production resource.
  
      
•     Select the resource in the table and select the "DELETE" button to remove the resource from the destination system. Since the initial sync never completed, this resource is not useful.
  
  
      
•     Navigate to the resource (volume or volume group) table, select the resource and then select the "MODIFY" button to launch the resource properties.   Deselect the checkbox for "Replication Destination" and apply the changes to allow production access to this resource.
  
  
  
  Remote System:
  
  The remote system can be removed after you have:
      
•     Unassigned the replicated policy from all the resources that were replicating from the failed source system, destroying the replication sessions.
      
•     Deleted any primary/user policies/replication rules on the destination system which were replicating back to the failed source system. You must first unassign the policy from the resources before the replication rules can be deleted.
  
  Navigate to the Protection->Remote System section and select the remote system in the table and then the "DELETE" button.    
   
  Note: Since the source system is no longer available, you will get the following error in the delete remote system job:         Failed to delete the remote system with ID from the remote side with IP . If the remote system is still present on the remote side, try to delete it locally from the remote side. (0xE02010020008)

   

  ---