Setting up a repository on ESXi

[johnzapf]

So I have a server with three RAID's. Raid one is running ESXI, RAID 2 is called DATA and it has all the VM's. RAID 3 is Storage.

I am wondering how to use RAID 3 on my Windows Server VM that's running Veeam without making it the D drive? Because if it I make it the D drive I can't back that server up as a VM because it has the storage on it and it will error. I want to make the storage repository independent "not as the D drive on the VM Veeam server?

So again I have the Veeam server running on RAID 2 and I want to set up the repository for all backups using raid 3 without mapping it as a D drive because then I can't back up the veeam server as a VM on raid 2, it will fail saying not enough room because it's trying to back up the D drive that 32TB's so I'm just trying to separate that and find out how to make the repository go directly to the raid 3 without using D drive on the veeam server?

[Mildur]

Hi John

We don‘t recommend to have the backup storage on the same hardware as your production VMs. This may be ok for a home lab.
For other use cases dedicated hardware should be used. Imagine the situation where your ESXi host cannot boot anymore because of a hardware issue.
You won‘t be able to restore your data until you have bought new hardware, reimported the raid configuration, installed ESXI again, build a new VM and imported all backups. Lot of time lost for a business until critical systems are running again.

Or worst, all physical disks gets destroyed somehow (water, electricity, theft, …). You loose your production VMs and backup data at the same moment.

If you want to go forward with your plan, deploy a new windows or Linux VM and use it as a dedicated backup repository server.
In the backup console you can add this server as a repository server:
https://helpcenter.veeam.com/docs/backu ... ml?ver=120

Best,
Fabian

[johnzapf]

Well thank you for getting back to me. In a perfect world that would be great. But I've been an IT guy for 34 years and have had hundreds of clients and not one of them wants to spend any money on IT. This project I'm working on now took me 3 years to convince them to move their old server to a new server. I had a budget of $2000 I bought a used Dell R740 and some new drives. That's the real world I'm just trying to make do with what I have.

[Sturniolo]

Hey there @johnzapf

Totally agree with what @Mildur says. That’s the way to go.

But I totally get it... sometimes things don’t get the budget or attention it needs...

One thing to possibly consider.. you could always use External Hard drives for a Repository and rotate them out. They might not be the fastest storage, but you will be able to successfully backup the workloads and have them off of the server.

Here is a link to our help center that talks about rotating external drives: https://helpcenter.veeam.com/docs/backu ... ml?ver=120

Keep in mind there are some limitations... *Listed on the link above*. But it might be a way to solve your issue.

Hope this helps.

[karsten123]

im sorry but this is total crap. that is not IT!

[tyler.jurgens]

If you have ability to deploy another VM, deploy it on the "RAID 3 Storage" and use it as a repository for your backups. Everything Mildur said is valid and you should avoid using the same hardware for the repository, but functionally another VM deployed to the other storage would work. You won't be able to add it to an existing VM in any fashion and make it functional.

Frankly, I would strongly heed the advice given above, but if you are completely stuck with what you have then at least understand the risks of doing so.

IT, unfortunately, isn't magical, and the saying "Just because you can, doesn't mean you should" rings true more often than not when people try to work around best practices.

[Origin 2000]

I am wondering how to use RAID 3 on my Windows Server VM that's running Veeam without making it the D drive? Because if it I make it the D drive I can't back that server up as a VM because it has the storage on it and it will error. I want to make the storage repository independent "not as the D drive on the VM Veeam server?

Thats a easy one.
1. Add another vDisk to your VBR VM and store this into the Datastore which runs on your 3rd. RAID
2. Format it with ReFS and the needed settings
3. Create another Veeam Repository on the server which already provide the default Repository
4. Move the existing Backups and change the Job settings (otherwise just create new jobs and use the new Repository... later remove the "old" backups)
4.a If you want the space back you need to run "sdelete <options>" and a svMotion to "shrink" the VMDK, otherwise just create a new VBR VM with the right size and install VBR again and import your config
5. Now the answer to your question.... if you would like to backup the VBR VM by it self you need to edit the "exclude" settings in the way that only your first vDisk aka drive c: is backup.

Another way to exclude a vDisk will be to adjust the "(in-)depended" settings within your VM configuration so that a VM snapshot isnt allowed anymore.

Regards,
Joerg

[COMO]

Like others are writing here, it is solvable.

But, sometimes working with IT we just have to tell our customer “No! We won’t do this for you because it’s too big of a risk and not a solution we’ll stand behind should it fail nor help you set up because it’s against all common practice”.

The main reason for this is of course securing the customers data but there is one more thing and that’s the simplest fact that if the setup/solution goes south, you can bet you ass they’re going to be pointing fingers at you and claim you’re the one to blaim as you helped set it up and managed it.

Don’t take me the wrong way here, the solution is creative, but for you own and you companys sake, don’t do it.

Sometimes customers need their IT partner to put their foot down for their own good as they clearly don’t know what they are doing and risking. This is definitely one of those times.
Feel free to show these posts to them as well to help them understand you’re not the only one having a bad feeling about this.

//Conny

[RobTurk]

Well thank you for getting back to me. In a perfect world that would be great. But I've been an IT guy for 34 years and have had hundreds of clients and not one of them wants to spend any money on IT. This project I'm working on now took me 3 years to convince them to move their old server to a new server. I had a budget of $2000 I bought a used Dell R740 and some new drives. That's the real world I'm just trying to make do with what I have.

If this is what you have to make due with, then that's what it is.
These internal disks can be their first line of defense for simple 'oops' situations. You may get away with it if you can persuade them to take a second line of defense.
Think 3-2-1 rule:
- Use the disks that you have to work with as local copy for data you want to be able restore
- Use an external storage location to create an off-site copy (backup copy, SOBR) for data you need to be able to restore

There are several per-TB storage providers (public cloud, dedicated storage clouds, Veeam Service providers) that can do the job for an off-site copy.
Many SMB customers balk at one-time investments but are OK with reasonable monthly charges. So sell it to your customer as a monthly insurance charge against disasters, ransomware etc.
Do explain that in case of a disaster, restores may be slower depending on internet bandwidth and other factors.

[18436572]

I had to do something similar. However it was with full understanding that it was a very bad practice. The backups were on the same SC Compellent array that the VM's were on. My justification was that we acquired new storage and these were to only get us through for a few months. Those backups are gone now...