Offsite storage

[brichardson1991_obk]

Hi all.

What's the best method of getting our backups/Replicas over to a offsite storage hosted at our DR site.

We've narrowed it down to 2 options

• Set up another replication job to replicate to the datastore at the DR site and just tell it for one host (we won't physically be using that host, we just want to store our data)
• Set up a backup copy job to copy our current backups to the datastore

The concerns we have are as follows:
A backup copy job will need to copy the entire job each night this is on average 70-100GB not including the full backup chain of 1.3TB (lots of bandwith i assume)
The Replication job will take an initial hit but be quicker after as only smaller changes, however we'd need to specify a host we want the replica's to fire up on and we won't actually be doing that.

The whole purpose of this is that we have a copy of the backups/replicas offsite so if we were without our office and can't function at the DR site, we can just pick up the storage unit and got to a Regus style office and start firing up VM's with replacement kit.

Anyone have any ideas which will be best? And how to go about doing this

[foggy]

A backup copy job will need to copy the entire job each night this is on average 70-100GB not including the full backup chain of 1.3TB (lots of bandwith i assume)

Not sure I understand this concern, since backup copy job is always incremental and copies only blocks changed since its previous sync to the target location.

[brichardson1991_obk]

So would the backup copy job need to copy the 1.3 TB every time it has a backup injected into it after the 14 days of restore points...
If so we cannot afford to push 1.3TB over the WAN on top of the incrementals.
Hence why we are looking into replicating them.

[foggy]

No, backup copy job will always copy only changes occurred since its previous copy cycle (even in case the latest restore point in the source chain is full).

[brichardson1991_obk]

so it will inject the increments on the target side regardless so only need to transfer the incremental file?

[chrisdearden]

That's not how it works. Think of the copy job as a separate backup job, that uses the primary backups instead of your source vm. As foggy says this means that the only data (beyond the initial full) that goes over is incremental. It's not a robo copy.

Sent from my A0001 using Tapatalk

[brichardson1991_obk]

So would we be able to recover from it if we lost our original backup.

Like the proposed solution. no access to office pick up the storage with the copy on it and restore from there?

Or would we be better using a replica?

[foggy]

Yes, if you have a full backup copy job chain available (initial full and subsequent increments), you will be able to restore to any of the restore points it contains.

[brichardson1991_obk]

but if we lost access to our original backup chain we wouldn't be able to recover using backup copy

[foggy]

No. Backup copy maintains a separate chain of backup files, you will be able to recover from it alone. Please review the user guide section I've referred to above, you'd get a better idea of how it works.

[brichardson1991_obk]

Thanks Alexander, I'll look into it further.

Just want to say thanks I've now read more and understood it.

This will work great for the use we want.