Data Domain: Replication Initialization is in Progress, and Upgrade Cannot Proceed

Symptom:

Error: Replication initialization is in progress, and upgrade cannot proceed.

Resolution:

1. Check if the Data Domain is configured for replication and if the replication is in an initializing state.

   #replication show config      ------> (To check if the replication pair has configured)
   
   #replication watch rctx://  ------> (Provide the context number to check the replication status)
    
        #replication watch rctx://1
        Use Control-C to stop monitoring.
        (0:00:00) Replication initialize started...
        (0:00:10) 1%: initializing 3/3^C

2. As the replication initialization is in a third phase, check the replication status. The user must wait until the replication initializations complete.

   # replication status rctx://1
    
        CTX: 1
        Mode: source
        Destination: mtree://ddve3010.dd.lab/data/col1/oracle-se2
        Enabled: yes
        Low bandwidth optimization: disabled
        Replication encryption: disabled
        Replication propagate-retention-lock: enabled
        Local filesystem status: enabled
        Connection: connected since Fri Oct 12 12:37:24
        State: initializing 3/3 1%       ---->  (Check how much the initialization completed)
        Error: no error
        Sync'ed-as-of time: -
        Current throttle: 8192K bps
        Max-repl-streams: 8 (default)

3. If the user wants to upgrade the Data Domain first, then replication must be broken on both sides.

   On source:
   # replication break rctx://<ctx_num>        ------> (Provide the context number to break the replication pair)
   
   On target:
   # replication break rctx://<ctx_num>        ------> (Provide the context number to break the replication pair)

4. After the replication is broken, upgrade the DDOS.
5. After the upgrade, replication can be reconfigured in one of two ways:
   • Scenario A: If initialization was getting close to finishing, fastcopying the destination mtree to a temporary mtree and redoing the replication may be best.
   • Scenario B: If initialization was not very far along and the progress can be discarded, delete the destination mtree and then redo the replication.

Scenario A: Saving replication progress

1. On the destination Data Domain, fastcopy the replication mtree to a temporary mtree and then delete the original mtree and run cleaning for one minute to clear it out:

   # mtree create /data/col1/<temp_mtree>
   # filesys fastcopy source <repl_mtree> destination <temp_mtree>
   # mtree delete /data/col1/<repl_mtree>
   # filesys clean start        -------> only needs to run for 1 min to clear out deleted mtree

2. Now remake the replication on the source and destination, and initialize it on the source:

   On target:
   # replication add source mtree://<dd_src_hostname>/data/col1/<source_mtree> destination mtree://<dd_dst_hostname>/data/col1/<dest_mtree>
   
   On source:
   # replication add source mtree://<dd_src_hostname>/data/col1/<source_mtree> destination mtree://<dd_dst_hostname>/data/col1/<dest_mtree>
   # replication init rctx://<ctx_num>

3. Once the initialization is done for the replication context, delete the temporary mtree off of the destination:

   # mtree delete /data/col1/<temp_mtree>

Scenario B: Starting over from scratch

1. On the destination Data Domain, delete the replication mtree, run cleaning, and then redo the replication:

   # mtree delete /data/col1/<repl_mtree>
   # filesys clean start        -------> only needs to run for 1 min to clear out deleted mtree

2. Redo the replication on the source and destination, and initialize it on source:

   On target:
   # replication add source mtree://<dd_src_hostname>/data/col1/<source_mtree> destination mtree://<dd_dst_hostname>/data/col1/<dest_mtree>
   
   On source:
   # replication add source mtree://<dd_src_hostname>/data/col1/<source_mtree> destination mtree://<dd_dst_hostname>/data/col1/<dest_mtree>
   # replication init rctx://<ctx_num>