Veeam B&R for DR/offsite

[techstarts]

I'm new to Veeam B&R product. I have been given a task to evaluate this product for DR
I think I understand backup and replication part.
But I cannot use replication, as we do not have DR site. We always hire a DR Site during our DR test.

So I'm wondering how I can use this product

For production site this is all I can think of implementing
1. Configure Veeam B&R on physical server.
2. Ensure it has HBA
3. Present all Virtual Infrastructure LUN to this Server
4. Also provide additional LUNS which will become backup destination for VM
5. Get these additional luns backed to the tape


During the DR, we go to DR site. However what to do for DR test is not clear to me.
Below is what I can think.

1. Restore the tape to LUN's
2. Rebuild ESX host
2. Present the LUN's to ESX host
3. Go and register each and every VM
4. Power ON the VM

So in DR site I cannot see the option of using Veeam B&R tool and it's features like instant recover.
Appreciate if someone can assist esp how Veeam B&R can be used for DR in absence of DR site

Appreciate if someone can point me to DR procedure in conjunction with Veeam B&R

Thank you,
Preetam

[J1mbo]

When Veeam B&R does it's job it will end up with a small number of large files on the backup server itself. It is these that are needed on tape or preferably USB disk, if they will fit. (As a side note - be VERY careful when presenting LUNs directly to Veeam, frankly this scares the **** out of me given the documented processes that can lead to Windows destroying the SAN file systems!).

At the DR site, the process would be get one or more ESX servers with sufficient disk space and preferably a second box running Veeam (maybe get this ready to go before hand, could just be a desktop PC really), hook up the USB drive, import the backups into Veeam, then restore them to the new ESX host(s). Then perform storage migrations to your production (DR) storage.

All of this will take time though, not least ordering hardware, if it is a regulated need it would be highly beneficial to get the tin in place before hand (IMO).

Hope that helps!

[techstarts]

When Veeam B&R does it's job it will end up with a small number of large files on the backup server itself. It is these that are needed on tape or preferably USB disk, if they will fit. (As a side note - be VERY careful when presenting LUNs directly to Veeam, frankly this scares the **** out of me given the documented processes that can lead to Windows destroying the SAN file systems!).

I will ensure this box has only read permission on these VM. I think this is the only option given the environment we have.

At the DR site, the process would be get one or more ESX servers with sufficient disk space and preferably a second box running Veeam (maybe get this ready to go before hand, could just be a desktop PC really), hook up the USB drive, import the backups into Veeam, then restore them to the new ESX host(s). Then perform storage migrations to your production (DR) storage.

Thanks so much, this part was bit hazy to me. Please help me understand why not directly restore them on production host?

[J1mbo]

this part was bit hazy to me. Please help me understand why not directly restore them on production host?

Because the VMs are stored by Veeam within it's own file format (on the Windows Server hosting Veeam Backup & Replication). This is because the data is de-duplicated and compressed right across the boundary of the job, so no sequence of blocks as they exist on the SAN will exist directly as they were on the WIndows server, let alone the files.

So when the new installation of Veeam is fired up, the first part of the process is to import the jobs - Veeam will scan through the files and work out what it has.

Next the VMs can be restored to the ESX host(s). The quickest way to do that is to run an instant restore, at which point Veeam is dynamically reassembling the VMDKs on-the-fly as they are being accessed by the hosts. Obviously this is CPU and disk intensive and hence slow. So, either use storage vMotion (if licensed) or a cold storage migration to move the VM then to 'production' (DR in this case) storage. This is the quickest way as there is a c.10MB/s limit in uploading files to vmware hosts otherwise.

Hope this helps!

[techstarts]

Thanks so much. I fail to understand why there is no document from Veeam on best practises. If it is some where I would like to read it.

[J1mbo]

But there are so many different options, and anyway there is this forum

Personally, I'm happy to have short (and in places sketchy) documentation if it keeps the price 10x lower than the legacy products like BackupExec..., they need a 300 page manual just to decipher the front screen of that.

[Vitaliy S.]

Hello Preetam,

Let me chime in with another advice. Since you're looking for best practices in DR scenarios, I'd suggest to verify most mission critical VMs with SureBackup jobs, this will guarantee that these VMs will be 100% recoverable from backup files.

Besides, if you search our forums you will see that most of the questions have already been answered: http://forums.veeam.com/search.php?st=p ... ackup+best

Hope this helps!

[techstarts]

Thanks for the tips.
To be honest Veeam B&R has lot of scope of documentation.
It is unfortunate people have to jump into forums to learn the product's best practices.

[J1mbo]

Certainly a knowledge base with decent search capabilities and deployment scenarios would be helpful. There isn't nearly enough available on the low-level stuff, for example information on the WAN aspects (packet sizes, TCP windows sizes, compression algorithms, and so on).