VEEAM backup concept in HO

[liviu.tutuianu]

Dear all,

We implemented VEEAM in many of our branches from all over the world. The branch infrastructure is very simple (only 2 ESX hosts that replicates their VMs over a 1 GB/s cross cable connection).

Very soon, we will replace TSM backup solution from each HO (Head Office) with VEEAM. As many users from this forum have good experience with VEEAM , I need your opinion on the concept that should be implemented in HO. We will create only replicas.

Let me present you the infrastructure from a small HO:

- 10 ESX Hosts with 4-5 VMs
- 1 VBR with 16 CPUs, 80 GB RAM, a Network Team configured of 4 x 1 GB /s
- Network switches of 1 GB/s
- 1 DataDomain Storage 38 TB connected to VBR over Fiber Channel. The replicas will be stored on this Storage.

I have in mind two concepts:

Concept 1. On each Hosts, I would create a proxy server. After that, for each replication jobs that will be created from that Host, we will use the proxy server from that Host as a source proxy and the VBR as a destination proxy. Each replication job will include one or more VMs.
Concept 2. We will not register any proxy servers and will let the VBR as Automatically proxy for each job.

For the first scenario, I am afraid that I will need to create a separate document for each host, so that, when the client will ask a new replication job, my colleagues will know exactly how to configure that job. The second one it would be more easy to be configured, but I am afraid of the performance impact. I think that the second concept would work great if we upgrade the network infrastructure from 1 GB/s to 10 GB/s, right?

So, what is your opinion? How did you proceed in your HOs? I would be very happy to get as many opinions as possible, so i can choose the best one that can be applied for us. Also, if you have some other concepts, please share it with us. Also, what monitoring solution for the daily jobs would you recommend? We were using VEM, but we have some bugs and we opened an incident some months ago and we were informed by VEEAM that it will be fixed in the next release...so, until then, we don't have a proactive monitoring tool.

Thanks you in advance,
Liviu

[foggy]

Liviu, could you please clarify, are you going to replicate between hosts in the particular HO? Not sure I understand the role of the DataDomain storage in your scenario.

[veremin]

Also, what monitoring solution for the daily jobs would you recommend?

Have you ever considered implementation of Veeam ONE, which can be integrated with VB&R, allowing you to keep track of data protection operations in your virtual environment and monitor performance of your backup infrastructure.

More information regarding it can be found in the corresponding user guide and at Veeam One forum.

Additionally, it might be worth taking a look at the existing topic at PowerShell subforum, which is primarily concerned with the reporting stuff.

Hope this helps.
Thanks.

[liviu.tutuianu]

Liviu, could you please clarify, are you going to replicate between hosts in the particular HO? Not sure I understand the role of the DataDomain storage in your scenario.

Dear foggy,

We will create CIFS or NFS on DataDomain and attach them to the same ESX host where the central VBR is located. On these, we will save the VMs replicas.

Have you ever considered implementation of Veeam ONE, which can be integrated with VB&R, allowing you to keep track of data protection operations in your virtual environment and monitor performance of your backup infrastructure.

Dear Eremin,

I thought that VEEAM One is used only for hardware monitoring. Am I wrong? I need a monitoring solution for VEEAM jobs. It would be great if it could send one e-mail/day with the backup status from last 24 hours, but the report should be more complex than the ones sent from VEM. The only solution that got in my mind is to create some Powershell scripts that would be scheduled locally on VBR.


Thanks for you both.

[kjc3303]

Hi

What storage are the prod vm's on? Will vbr server have access to the storage that the prod vms is on?
Which esxi host will the replica vm's be hosted on?
Will they be on the Same esxi server as the vbr server? Or will they be replicated between each other?

Kev

[foggy]

Anyway, you'd better have virtual proxies on both source and target hosts to use hotadd to retrieve data from source and populate it on target.

I thought that VEEAM One is used only for hardware monitoring. Am I wrong? I need a monitoring solution for VEEAM jobs.

Starting from v6.5, Veeam ONE allows to monitor the health and performance of the Veeam B&R infrastructure and provides prebuilt alarms for monitoring its metrics and events. In other words, both products are deeply integrated.

[Vitaliy S.]

It would be great if it could send one e-mail/day with the backup status from last 24 hours, but the report should be more complex than the ones sent from VEM.

This report is going to be part of v7 release, meanwhile you can use Protected VMs and Weekly VM Backup status reports that are already available in Veeam ONE v6.5

[liviu.tutuianu]

Hi

What storage are the prod vm's on? Will vbr server have access to the storage that the prod vms is on?
Which esxi host will the replica vm's be hosted on?
Will they be on the Same esxi server as the vbr server? Or will they be replicated between each other?

Kev

Dear kjc3303,

The prod VM's are stored on the ESX hosts from HO. From, there, using VEEAM, over the internal network, they are replicated on DataDomain storage. Basically those LUNs from the DataDomain are mounted to the VBR.

Yes,the VBR has access to all the ESXs where the VMs are stored on. ALso, it will have access to the destination (in our case, DataDomain system).

[yizhar]

Hi.

A question to liviu.tutuianu:

Why do you plan to use only Replica and not Backup?

Replica is good for quick recovery (short RTO and RPO), but with only few restore points.
Backup is better for recovery from several points in time, with better deduplication as well.

I suggest that you also configure a weekly backup job, in addition to the replica jobs.

Yizhar

[vmJoe]

Hi Liviu,

I have a few customers that replicate data across the USA and other parts of the world using the process described in your option 1. I think that this offers the best mix of performance and flexibility in deploying across a LAN/Wan infrastructure.

The documentation for this this process seems like it should be straight forward and easy to reproduce.

Keep us posted on the option you choose.

[liviu.tutuianu]

Hi.

A question to liviu.tutuianu:

Why do you plan to use only Replica and not Backup?

Replica is good for quick recovery (short RTO and RPO), but with only few restore points.
Backup is better for recovery from several points in time, with better deduplication as well.

I suggest that you also configure a weekly backup job, in addition to the replica jobs.

Yizhar

Yes, we will use backup jobs for some of the servers and replication jobs only for servers that are very critical and that don't have more than 100 GB HDD.

Regarding the long-term data, I think we will use another software and save data directly on tapes.

[veremin]

I think we will use another software and save data directly on tapes.

As an additional option you can also utilize PowerShell script and specify it as post-job activity or set it to run at any given time via Windows Scheduler.

Hope this helps.
Thanks.