Best recommandations for DR site for backup/replication (offsite)

[Simon_LBC]

Hi everyone,

My current setup for local backup & replication are pretty simple; I have two local 2012 R2 Hyper-V servers with two separate direct attached storage (one for each), the first server is production, the second server is a failover/replica. Every night (24h) I have a replication job between the two servers (with 2 incremental checkpoints) and right after the daily replication job is completed, I have a backup job for the same VM/data, where the replica server/storage is also set as backup repository, where I keep a 30 days incremental backup for data retention, in the case of accidental data deletion by my users.

Yesterday, I have ordered a third server that I want to setup in a remote DR site (offsite) and where I also want to set a replication job and a backup job for data retention, exactly the same that local, but on a offsite location. This new server will simply use a large built-in storage instead of any external storage, but this doesn't really matter... the remote DR site will communicate with my local network using a hardware (physical) site-to-site VPN upstream to the DR server.

The initial (full) replication job with the new server, SO AS ,the initial (full) backup job will be performed locally before physically moving the new server to the DR site. Despite that our total data size is pretty huge (around 30 TB before compression/optimization), our daily change rate is VERY limited, only around 10 GB to 20 GB per daily cycle (after data reduction). For this reason, the internet connection between the local network and the DR site will be pretty limited; a 100/100 Mbps fiber connection. But if we need to go live on the DR site for some reason, our hosting provider can perform an immediate upgrade to a 1000/1000 Mbps connection of even more.

I have read a lot of documentation from Veeam KB, however I am not quite sure how to integrate this new backup & replication offsite destination into my local daily routine;

• Is it better to use a "backup copy" job from local backup job to the DR site or to create a separate job for backup from local to DR site?

• Is there any "replication job copy" that I can use for my DR site or replication is only available as a second separate job?

• I also read about "Replica Seeding" option for low bandwidth WAN connection, but I am not sure to understand how it work and if it will be gainful for me for my DR site?

BY THE WAY; my Veeam licence is the Standard so I can't use any WAN Accelerator features and I don't plan to upgrade to higher licence because of my very low data quantity I had to upload daily to the DR site. I see no return on investment by purchasing a licence upgrade.

Thank you very much for your recommandations!

[DGrinev]

Hi Simon,

1. It's better to use backup copy job from local to the DR site, since it takes VMs data blocks from the local repository without impact on the production VMs.

2. Replication is the only way as a second separate job.

3. Replica Seed will reduce the amount of transferred data during the initial Replication job. It will compare the difference in VMs data blocks contained in the "seed" with the source, thus, it eliminates the need to transfer all the data blocks. It's recommended approach for starting replication to the DR site, especially when you're going to replicate 30TB VMs. Thanks!

[Simon_LBC]

Hi Simon,

1. It's better to use backup copy job from local to the DR site, since it takes VMs data blocks from the local repository without impact on the production VMs.

Sounds good!

2. Replication is the only way as a second separate job.

Got it! Sounds good!

3. Replica Seed will reduce the amount of transferred data during the initial Replication job. It will compare the difference in VMs data blocks contained in the "seed" with the source, thus, it eliminates the need to transfer all the data blocks. It's recommended approach for starting replication to the DR site, especially when you're going to replicate 30TB VMs. Thanks!

I understand that the "Replica Seed" option is only available for the initial (full) replication job, right? So, since I'll make the initial replication job on my local network (on a fast LAN connection) before moving the server to the DR site, I assume that this option is probably not really relevant, except maybe to shorten the initial replication (full) job time over local network?


Thank you again!!! I love you clear answer

[DGrinev]

3. That's correct. In this case you can use replica mapping after the new host will be relocated to the DR site. Thanks!

[Simon_LBC]

3. That's correct. In this case you can use replica mapping after the new host will be relocated to the DR site. Thanks!

Hi again,

I'm sorry but I'm not sure to understand that part. If I take my new server, connect it temporarily into my local network, create a new (second) replication job, run the initial replication (full), then finally simply move the server into the DR site which is connected using a site-to-site VPN, what's making the remote network connection 100% transparent for the host, why do I need to re-map my VMs thereafter?

Or do you mean that the replica mapping will be required if I use the replica seed features for the initial job?

Thank you!

[DGrinev]

You don't need to re-map if the host will be recognized the same after the relocation, the replica mapping will be required only if the target host will get a new config at the DR site. Thanks!

[Simon_LBC]

Got it! Thanks!

[Simon_LBC]

1. It's better to use backup copy job from local to the DR site, since it takes VMs data blocks from the local repository without impact on the production VMs.

Hi DGrinev, I just put my hand on my new DR server and storage and I am now setting up everything. I am wondering if I setup a "backup copy" job from my existing local job (that is in place for several months/years) at this point is this will only copy the last incremental backup or somehow it'll copy everything that is related to this job and that is present in my current backup repository??

[DGrinev]

Hey Simon,

The backup copy job will create a full backup file with the most recent state of VMs available in the local repository (The list of VMs according to the source job). Thanks!