Case ID: 5186685 Veeam configuration

[hollow]

Dear all,

I am new to Veeam and I would like to find out if the following is possible with Veeam or not.


Scenario: (Please refer to image @ bottom of post)

My customer has multiple ESX production servers @ site A.
The proposed solution is to have an IBM x 3650 M3 server, with 2.4TB local storage as a Veeam backup server to backup the ESX hosts.
After backing up to the IBM x3650 M3 server, it is to replicate its backup to another Veeam server @ DR Site @ another location via WAN.

I would like to know the following:

1. Can this be done?
2. If yes, how? Able to provide step by step illustrations?

My customer already has purchased Veeam v6 but it is sitting there doing nothing.

[foggy]

Hello! This scenario is definitely possible, please search this forum for numerous existing discussions on how to perform offsite replication of backups. Here is the couple of the most recent ones: Best backup setup for mirroring backup respository and v6 - How to have a local and off-site backup copy?

[hollow]

Thanks foggy for the advice.

Is there a guide on the following:
1. how to back to local > I played with it and figures out this
2. how to replicate from local veeam server to off site veeam served.

number 2 is my main problem

[Jfmoots]

Hollow,

v6 does not copy/replicate it's backups from an onsite to an offsite location. You could accomplish this by creating two jobs per vm/group of vms. Job 1 would target the local repository, Job 2 would target the remote repository. Both jobs would leverage CBT and will not interere with each other's ability to leverage CBT. Another popular way to get this done is by using 3rd party tools to sync folders/files (Rsync comes to mind) The links that foggy posted are great discussions about a variety of ways to accomplish this.

[hollow]

I have got this from Veeam's FAQ:

Q: What are Replication jobs designed for?
A: Replication jobs produce exact replicas of production VMs on standby hosts. These replicas can be powered on immediately when production VM goes down, without any dependencies on Veeam Backup and Replication server and at full I/O performance. However, replicas require standby host, and much more disk space due to being uncompressed. Thus, replica are typically are only used for tier 1 VMs with low recovery time objectives.

My offsite is an ESX host not a repository.

So in the case, can my scenario work?
1. Backup to local veeam server.
2. Replicate from local veeam server to Off site stand by ESX hosts.

[foggy]

Replication copies a VM from the source HOST where VM resides (not from the backup repository) to a destination HOST. To have both local backups and offsite replica of your VMs, you need to create one backup job, pointing to local repository, and one replication job, pointing to the offsite ESX.

[Fiskepudding]

my 2 cents:

Your offiste is an ESX host you say. But does it NEED to be?
(if you want Veeam Replication offsite, yes it has to be and HOST)
But you can always install openfiler or something on it and make it a repository?

You mention nothing about the bandwidth of your WAN.
The main reason you want a replication job is to recover fast from a disaster. (No need for restore large VMs)
So in case of disaster you can just power on these VM's that you have replicated.
However, if this is supposed to work, the WAN link need to be sized to its need. Since you wil end up running production traffic over WAN.
Depending on the wan, or the distance for that matter, (If it’s a 5 min drive to pick the ESX up) you should think twice about where you have your replicas and where you have your backups.

[hollow]

Thanks for the advice. Originally my idea is to backup the production vm to a Veeam server with 2 NIC cards, 4 ports (2 NIC teamed to production, 2 NIC teamed to replication) (see diagram below) and then replicate the backup from the veeam backup server to the off site DR ESX host, also a Veeam backup server (VM) via a 100Mbps WAN (missed that info out on the diagram above)

As foggy mentioned, replication copies from the source host and not able to replicate from the Veeam server. So design wise, I will need to change.

I am thinking how to change the design to fit into my requirements. Eg: Do a scheduled copy from my veeam backup server to off site but dont think this is a good idea. Any experienced guys here?

* I have edited my .jpg file below.. hopefully it is clearer.

Edit: Unable to edit my .jpg above. I'll repost it:

[Fiskepudding]

I migt start by adding that I am far from a Veeam expert, I will anyway put down a few more words.
Hopefully some experts can and will correct.

100Mbps is very good for WAN. But keep in mind that bandwidth available for all your VMs in production 3Gbit (1+1+1).
If the average network usage is above 100Mbps, you will run into problems if you need to run off the replicated VMs, regardless of great WAN speed. An if it is an VPN link there can be much overhead.
Remer you have 30 times less bandwidth. It’s just math...you probably get the picture.

As far as I can say a scheduled copy of the backups from Veeam backup server to offsite (ESX) would do.
You would however need some sort of target to store them on. Typically a NAS application on a VM., if you don’t have anything else on that site.
Open filer has embedded rsync server, witch I have used several times with great success.

To copy the files FROM the backup server (onsite), rsync is already mentioned, you also have Robocopy witch is aslo quite stable.

We use robocopy to offload local backup files to external HDD for archiving offsite.
This way can have as many restore points as we are willing to buy cheap SATA disks.
If I understand correctly you are moving away from replication then?

If are not moving away from replication, you don’t need to change much, if you just do a backup to local repository and do a replication to the host on the WAN side.
This requires that 100Mbps is enough for production in case of a disaster. Unless you want to pick up the host and bring it on site...

We backup and replicate to a “local DR site” if that’s a word…This is done over 10Gbit network quite a distance from our production environment, and many fire zones away.
Off cause if your entire premises is "nuked" we have to rely on the offsite backups. But hey.. I guess we have bigger concerns then... Like workers and having a place to put a new production environment.
We have found this acceptable, and we estimated that our IT infrastructure and data would be up and running long before the rest our business (production) .. as always"your mileage may vary"

[dellock6]

Johnson, based on your design the Veeam proxy with 2.4Tb is in the local site (I must suppose it's also a repository), while Veeam Server is directly in the DR site, installed in a VM inside the ESXi, right?

In one of your first posts, your concern was to create a copy of the backups saved at local site also to the DR site. You basically have two way:

- if you can sustain two backup activities from the same VM = 1 backup job and 1 replica job of the same VM
- if you want to do a single job on any VM = 1 backup job to local repository, and then add a nfs/cifs storage in DR. With tools like rsync (there is a thread about this topic) you replicate backup files to DR storage. From here, you can import backups in the Veeam Backup (it's already in DR), power on the VMs via Instant Recovery in a short time, and them move the VM disks inside ESXi via storage vmotion (if you have the license) or via Veeam QuickMigrate.

Luca.

[hollow]

Thanks all. I will try with the suggestions given. Will post if I encountered any issues. :DDD