Break and Resync Replication
PURPOSE
This article lists steps on how to perform a replication resynchronization (resync) procedure on a system configured for directory/pool replication. For a more in-depth explanation of the purpose of a resync procedure, read here Data Domain - Replication Resync Explanation.
Here is a list of some reasons a break and resync procedure may need to be performed:
APPLIES TO
SOLUTION
The steps listed in this solution assume a basic replication configuration where a source Data Domain is configured to write to a destination Data Domain. Replication configurations can be much more complex consisting of bi-directional replication and chained configurations involving multiple systems. If a context is broken, the partner system must also have the related context broken. The break and resynchronization procedure requires that the file system to be temporarily unavailable for all systems involved.
Note: Do not disable any unrelated replication contexts as they will recover automatically when the file system is re-enabled.
This procedure can be performed using the Enterprise Manager UI (Option 1) OR the command-line (Option 2).
The EM process is simpler, and can be completed with a few clicks.
Break the existing replication and resync:
1. Enterprise Manager option:
a. Break existing replication
This removes the context from both the source and destination DDRs.
b. Resync the context
This adds the context back to both the source and destination DDRs and start the resync process. The resync process can between several hours and several days, depending on the size of the system and current load factors.
-- or --
2. Command-line manual break and resync option:
# replication show config CTX Source Destination Connection Host and Port Enabled --- ------------------------------------------------------ ---------------------------------------------- ------------------------------------ ------- 1 dir://test.datadomain.com/backup/monthly dir://test2.datadomain.com/backup/monthlytest test.datadomain.com (default) yes 2 dir://test.datadomain.com/backup/weekly dir://test2.datadomain.com/backup/weeklytest test.datadomain.com (default) yes --- ------------------------------------------------------ ---------------------------------------------- ------------------------------------ -------
Copy both outputs temporarily to a text editor (such as notepad) and label each section appropriately. This information will be needed when recreating the replication contexts. Note context numbers if targeting specific contexts.
(Ensure that backups have been paused (or any other systems using the Data Domain) and cleaning is not running).
# system show stats 07/15 15:04:40 INTERVAL: 5 secs -------------------------------------------------------------- CPU FS FS Net MB/s Disk KiB/s Disk NVRAM Repl busy ops/s proc in out read write busy KiB/s KB/s ---- ----- ---- ----- ----- ------- ------- ---- ------ ------ 0% 0 0% 0 0 0 0 0% 0 0
The system should show as idle.
# filesys clean status Cleaning finished at 2012/04/10 06:01:30
Cleaning should indicate finished.
d. Disable the filesystem on the source system, if systems are running DDOS version 4.7 and below. (Bug 6995 for more information). This is an optional step for DDOS version 4.8 and above.
# filesys disable
e. Break the replication on the source system.
# replication break rctx://<context number> Example: # replication break rctx://2
# replication break context all
# filesys disable
# replication break rctx://<context number>
# filesys enable
i. Recreate replication contexts on source and destination system. Use the previously saved configuration information to rebuild the deleted replication contexts.
# replication add source dir://<source directory> destination dir://<destination directory>
Tip: Copy the command from the source clipboard and paste it in the destination clipboard as same command is used on both source and destination systems.
# replication resync rctx://<context number>
Note: This command is run on the source system only.
# replication watch rctx://<context number> # replication show performance [all | rctx://<context number>] # iostat 2 (CTRL+C to quit)
Note: The iostat command displays a wider system view. The display window may need to be maximized to prevent the text from wrapping and distorting the output.
If the source directory is empty, the resynchronization will occur quickly, and may not be represented in the output. In addition, if no data is being written to the source directory throughput numbers will be low (or non-existent).
Note: From DDOS version 5.1, Pool replication is not available, it changed to Mtree Replication.
If your problem persists after performing the steps in this article, contact your contracted support provider or Create a service request and Upload a Support Bundle (SUB) Data Domain: How to collect/upload a support bundle (SUB) from a Data Domain Restorer (DDR)
REFERENCES