Veeam – Backup VMs in remote sites

Reading Time: 6 minutes

I was wondering why I haven’t talked about Veeam when I use it almost every day in my job, not only administering backups but doing new implementations.

Recently, I had to implement a design where I need to backup VMs in remote sites but not back up them in a centralized storage, they will be backed up on each remote site storage.

So, by deploying a VM with the Backup proxy service and also use it as the backup repository we can accomplish the goal. We will save bandwidth and increase the speed to restore and backup those remote virtual machines by using the local storage on each remote site.



The scenario I am talking is the following, a dedicated VM with Windows Server 2016 Standard (a.k.a. W2016 STD) to act as a backup proxy and backup repository and Veeam B&R installed on the main site (the cloud we will say).

This is the high level design:


So, we are going to back up all the VMs that are hosted in the remote ESXi hosts and also save the backup data in the local storage.

As said before, in this way we save bandwidth and gain speed in the backup and restore process in case we need to perform any of it.

We will assume that we have a vCenter deployed with Veeam B&R installed. The Veeam B&R has configured the vCenter and then all remote ESXi hosts.



The implementation is pretty straightforward, we will have a dedicated VM to be deployed on each remote site and then perform the following high-level steps:

– As a backup repository, we are going to add a hard disk to the remote VM and use that hard disk as the backup repository for the site. We will seize the capabilities of Windows Server 2016 and use ReFS as the filesystem for the added hard disk.

– Install a backup proxy service, we just need to deploy the backup proxy service from the Veeam B&R console to the VM that we are using. The backup proxy will be who processes jobs and delivers backup traffic.

So, let’s go each step!


Backup proxy service

First, our Windows Guest OS VM is joined to the domain, so we won’t have any kind of problem for resolving the name or accessing with domain account credentials.

Let’s add the proxy by going to the Backup Infrastructure tab > Backup Proxy > Add VMware Backup Proxy…

As this is a new server for Veeam, we will have to add it as a “server” by pressing “Add New…”:

Then, this window will appear, just enter the FQDN of your server:

Choose credentials and chooseApply “to install the transport service:

After that, you will be able to choose the newly added server ( from the drop-down menu:

Now, let’s configure the Transport mode and Datastores for this proxy (as in the previous screenshot):

And for the datastores, choose the ones that are connected to the ESXi host where the VM is hosted by selecting Manual Selection and adding them:


After configuring that, you will have the same configuration as in this screenshot:

Finally, just hit Next and apply any kind of traffic rule if you want:

Now, finish, and the proxy will be fully configured and ready.



We configured these options because they are the best for our deployment which is using a Windows VM that will have a backup repository which will save the backups.

For more detailed options about the Backup Proxy service go here.

After configuring each backup proxy we will have a bunch of them in the Backup Proxies tab:


Backup repository configuration

In this step, I suggest following this article to perform this step.

Basically, we just have to add a new hard disk to our dedicated VM as Thick Provision Eager Zeroed, format the disk as ReFS and finally, add the Backup Repository to the Veeam B&R Console.

In that article, it’s also explained the benefits of ReFS so, I think it’s more detailed and easy to follow it.

After we configure all the backup repositories, we will have the same amount as the backup proxies:


As you can see in the previous screenshot, the path (D:\Backups) is the disk that we added to the VM on each remote site. We have configured the backup repository to that path because, as explained before, we have a disk formatted in ReFS and it’s explained in the article.

Backup job configuration

After configuring the backup proxy and backup repositories on each site, we are ready to the last step, configuring the backup job to perform backups.

Go to Home tab and then Backup… Virtual Machine:

Now, step by step, pick a name for the job:

Proceed to select the VMs you want to backup (in our case the ones in the EUR site):


Let’s continue and in the Backup proxy, click Choose… and select the correspondent backup proxy (EUR_proxy):


Press OK and go to Advanced. Configure it like that if you want Synthetic  full backups:


And then the monthly health check (recommended):

Accept and here is the summary for the backup proxy step (we will keep 7 restore points in our case):

Configure any option as you like (not in my case):

And finally, proceed with the schedule that you want after finishing the configuration for this job!

And that would be all for this remote site. We had to to the same with the other remote sites and our job will be done!



Finally, with this design you will be able to back up remote sites and store the backups in the local storage from each site.

If you don’t want to use a dedicated VM as a backup proxy, you can install the service on a VM that has low usage and install the backup proxy service, however, it’s recommended to use a dedicated VM which will have the backup proxy service and the backup repository (the virtual hard disk attached).



Cohesity Build Day Live

Reading Time: 8 minutes

I am going to share some thoughts and opinions about a recent video from the Cohesity Build Day Live recorded recently with the Build Day Live! team.

Disclosure: This post is sponsored by Cohesity.

First, just let me introduce briefly you what is Cohesity:

Cohesity is a platform focused on managing and consolidating secondary applications and the data. It provides a unified view and access to converge that secondary data, such as system backups and analytics in a simpler way to an IT administrator.


Now, let’s deep into the topic.

In the video, you will see how Alastair Cooke and Bharath Nagaraj building a Cohesity cluster from the scratch, configuring jobs, updating the physical appliance, restoring some data and showing some other cool stuff.

I really like this kind of videos because, you can see how they install a cluster, configure it or resolve any problems that can happen in real time without cuts.

Also, you can notice how much time it can take to deploy and configure a Cohesity cluster in some minutes, or even upgrading the whole cluster (node by node) while running some protection jobs (backup jobs).


In this case, they use a  physical unit for deploying their solution, so it’s a 2U enclosure with 4 servers/nodes inside (blade server type).

It comes, like most other solutions, with 1GB ports for Management purposes and 10GB ports for Production Data.

As this is an HCI solution, it comes with the storage and computes resources necessary to process and store all data (PCIe Flash card and hard drives in each node).


Cluster configuration and UI

To configure the cluster you won’t need a lot of data to fill or knowledge to do it, they configure the cluster easier than I thought and straightforward.

In a real scenario, a Cohesity engineer will do it for you thus, this is just to let you know of the simplicity of it.

The UI is simple and clean, the home dashboard looks nice with some graphics regarding your Storage Reduction, Health, Throughput, Protection Runs, etc.


As you probably guess, it backups your vSphere/Hyper-V/Nutanix environment like other products do, so you can configure a Backup Policy with a schedule, time retention, etc. to back up your data and then you configure a Protection job which will be the backup job that is associated to a policy.

Just register the hypervisor of your choice and basically, you’re ready to back up your virtual servers (VMs).

One option I really liked when registering a hypervisor was, the option of selecting “Auto Cancel Backup if Datastore is running low on space”, so the DataPlatform solution is aware of the datastore’s space and can avoid you a big problem there…


About granularity, there is a lot of options to select when you create a protection job (DB, Virtual/Physical Servers,  but regarding what you can see in the video they protected only VMs and Office 365 mailboxes in different backup policies.

It’s great that when you are creating a protection job (a.k.a. backup job). you can select an object like a cluster or a folder with some particular VMs and then check the “Autoprotect” option to ensure that new VMs that are added to that object (folder, cluster, etc.) will be automatically protected.


Regarding long term retentions, you can choose to add an external target like a NAS or any cloud (AWS, Azure, GCP, etc.) to store your archive backups there.

This is an option that adds great value to your strategy because when storing great amounts of data for several years, you usually don’t want to store it locally or even in a NAS.

In my opinion, having a flexible option to store it in any cloud can save you a lot of headaches despite the money that you must pay for the cloud service (which nowadays almost every company does).

So, within a Backup policy select the Archival option and then you add as many external targets to store your long term backups.



Your backup strategy is useless unless you can restore from it…

I do like some points about this section that makes so simple to restore, from a single file (even download it) to restore tons of virtual machines to your virtual environment.

– Single file restore

If you want to recover a file, you don’t have to search for the date, where it was, etc. As simple as searching for the name of the file (or the portion you remember) and it will be searched in the entire cluster for you:

And then, when you found that file, look at the options that we have:

First, search for the date and then, you can choose the usual option (Recover to Server) or … download it at the moment (a cool option there).

It looks like a painless and simple way to restore files that probably a non-tech person could do.

– Instant mass restore

Now, going bigger, let’s talk about the Cohesity instant mass restore of virtual machines. As the Cohesity platform is designed in a distributed architecture where there isn’t a centralized bottleneck, they can restore tons of VMs quite faster than other products.

When recovering a lot of VMs, in the background (you could look at your vSphere environment) it will mount an NFS datastore and bring up all you requested VMs (quite fast to be honest).

– Office 365

Finally, the last thing to show you is the option to backup your Office 365 environment. You can integrate Cohesity with your Office 365 and perform protection jobs that will be associated with a policy and consolidate all the data within the same platform.


The process is straightforward, selecting a package from your local computer or getting it from the Internet, this makes it so easy to do it for yourself.

One thing that stuck in my mind was that, while there were running some protection jobs you are able to upgrade the whole cluster (node by node) non-disruptively.

As the entire solution is designed to tolerate one node failure (N+1 redundancy) thus, you can upgrade one node without disruption in the service.

As we said before, the Cohesity platform is based in a distributed architecture so, in case a reboot is required after upgrading one of the nodes, you will only lose the bandwidth coming from that node and not impacting the rest of the environment.


Cohesity Helios is the console that lets you manage and view all your clusters from one console. As it’s in the cloud, you only have to register your Cohesity appliance and at the end, it will show up in the Helios console.

Helios Dashboard is similar to a Cohesity management dashboard but with the ability to manage all your clusters from that single pane of glass.

And it’s just not that… Helios lets you install applications!

Yes, you can choose to install applications in one of your clusters without anything else. What Helios will do is to deploy an app within a container (using Kubernetes) in that cluster without having to worry about the underlying infrastructure.

Just install, configure and run your app (as it sounds).

For example, running Splunk to gather data analytics in your clusters without having to worry about to deploy it is really a nice feature to look at it. 

I’ve never seen a feature like that and it really surprised me when I saw it. A nice additional value that you can consider when using Helios with your Cohesity platform.

Other use cases

As the Cohesity platform is cloud and hypervisor agnostic, you can protect objects on any cloud Azure/GCP/AWS or any hypervisor Hyper-V/VMware/Nutanix but, do you imagine what else can you do?

Well, you can use it to migrate VMs between different environments! It’s a great use case where you can choose to backup all your vSphere environment and move it to Nutanix for example or moving it to Azure.

Obviously, there is work to do after it but, the amount of simplicity that gives you with that, for me, it’s massive.

That’s all…

We saw a lot of things from the Cohesity platform, how can help your company to achieve that data consolidation by: backing up from different clouds and environments (cloud and hypervisor agnostic) , establishing an SLA in your services (configuring policies), recovering tons of VMs and other features like Helios, a cloud console that brings you a unified view for your Cohesity environment, analytics for all your data and even the ability to deploy applications without needing any kind of resources.

If you are interested in more content, check the Cohesity Build Day Live web page or the official web page from Cohesity.