Celerra Replication - Using fs_replicate against brand new file systems

Question

I have experience setting up Celerra Replicator against file systems already in production, but was wondering if I could take a few shotcuts against brand new file systems which are not in use yet. Is it necessary to complete all 9 tasks as documented in the "Using Celerra Replicator" against a brand new empty source file system?

Also, is quota info propagated to the file system on the target side, or do I have to set up quotas manually prior to replication on the taget side?

Thanks in advance for any responses.....

RApel1 · Answer

I understand the need to use fs_copy to copy the metadata, but there is more than likely no need for the incremental portion on a file system which is not in use yet. I would think that fs_copy the first time, then start up replication and be done with it. The question remains as to the details of whether or not you leave the file in raw of uxfs format after the first fs_copy.

Rainer_EMC · Answer

yes, since even with an empty file system we need to replicate the bits ( acouple of percent) that contain metadata like inodes.

BUT the fs_copy commands will obviously go much faster.

If you want it simpler I highly recommend to use Celerra Manager - it will do all these steps in one go. All you need to do to create the destination fs and then a couple of mouseclicks to start the replication.

Then just sit back and watch till task status tells you its done.

yes, quota info is replicated as well.

Rainer_EMC · Answer

obviously out of the nine steps the first two you only ever need to do once for a src/dst pair and not per file system

1. Establish communication between the replication sites.
2. Verify the communication link.

3. Create a SnapSure checkpoint of
4. Create the destination file system on the destination site.
5. Copy the baseline checkpoint to´the destination file system.
6. Begin the replication process.
7. Create a second checkpoint of the source file system.
8. Copy the changes between the first and second checkpoint.

9. Check the replication status.

nandas · Answer

I agree with you - if there is no data at all, there is no need to take the second checkpoint and run the incremental fs_copy. However, there is no harm in doing the same as a standard practice as there will be no incremental data to be copied and the process will be complete soon.

However, if you are not doing the second checkpoint and incremental fs_copy, there is no need to use -option convert=no in the first fs_copy command. You may either use Yes option or don't use any option (by default fs_copy will convert the destination rawfs to uxfs).

Others may provide different views though.

Thanks,
Sandip

Celerra

Celerra Replication - Using fs_replicate against brand new file systems

Was this post helpful?