Initial sync between sites

[Stefano Colombo]

We're setting up a remote replication between sites

In one site we have a cluster of two 4330 and in the recovery site a cluster with VSA .

Since the volume are quite big , and it would take ages to do the first replica via WAN , we followed a LH doc which explained to create a temporary MG and use it to move the first replica from site A to site B

The document then says to create a snapshot in site B from the temporary MG and finally to start a scheduled replica from site A to site B

This should use the initial snapshot and incrementally update it , but it looks like as it start from scratch .

has anyone tried , and succeed , to do the same ?

Is there any update documentation on how to do it ?

thanks

[oikjn]

assuming you have the hardware to spare, the thing to do is create a VSA joined to the DR management group in its own cluster, but located physically at the primary site.  Then create your remote snapshot to that cluster.  Then you can power down that VSA, ship it to the remote site and power it back up and as long as the networking was setup correctly, it will resume its operation as normal and you can do future snapshots without having to worry about the initial remote bandwidth needs.

If you are trying to move this to an existing cluster in a remote management group, you can still do this as long as you have enough extra storage to create another temporary VSA which you can use to make its own cluster and then when its back at the remote site you can then do a LUN migration to move the LUN from the temporary cluster to the final one.

[Stefano Colombo]

hello Oikin
the first method is what we should have done if we could at first install the secondary storage at the primary site . But it wasn't an option

As per the second method
is what we did , we create a temporary cluster and a remote snapshot on this
But we did not use a move to the remote site's cluster but instead a remote snapshot as suggested by Lefthand documentation

In any case the important part I suspect is how creating the final remote snapshot schedule from primary site to the recovery site will use the moved snapshot as base for incremental updates