Solved

Veeam backup to tape design


Userlevel 5
Badge

Hello,

 

Customer have an existing veeam on prem setup plus tape, they do manually  tape vaulting - they manually transfer the tape from HO to dr site (vaulting site).

What is the best and seamless approach so that customer no need to go to Dr site to do vaulting.

Can we relocate the Tape from HO to Dr, setup veeam resources (server and storage) and have copy/move from HO to Dr then do the tape backup jobs? 

 

Thanks

icon

Best answer by JMeixner 7 September 2022, 09:55

View original

12 comments

Userlevel 7
Badge +7

Hello,

your customer does backup copy to tape and not primary backup? Is the network connection stable and has it enough bandwidth to send the backups to the DR site?

If both answers are yes, then you can move your tape library to the DR site and sent your backup copies to the DR site directly.

Userlevel 7
Badge +3

Hi,

Yes that is possible. Depending on your setup, what you could do is copy your backups to the DR site and then run the backup to tape job there. 

Another option is that you could perform a Veeam backup to a repository at the DR site and then run the tape backup job there. However, do bear in mind that if you setup another Veeam instance, you may need additional licensing.

Userlevel 5
Badge

Hello,

your customer does backup copy to tape and not primary backup? Is the network connection stable and has it enough bandwidth to send the backups to the DR site?

If both answers are yes, then you can move your tape library to the DR site and sent your backup copies to the DR site directly.

they have backup to disk, then disk to tape. Both are located within the site

Userlevel 7
Badge +7

OK.

you will need a server at the DR site as a Veeam tape server and connect the tape library to it.

The network connection between the primary site and the DR site has to be stable and has to have enough bandwidth to send your amount of data in your time window.

The scenario is a standard scenario, no problem.

But please keep in mind that you make a configuration backup of your VBR database and copy it to the DR site, too. With this you can setup a VBR server at the DR site and restore the backups.

 

On the other hand, you can move the VBR server to the DR site, too and keep a simple repository server at the primary site for your primary backups. Then you have complete backup environment at your DR site in a disaster. But here you should copy a configuration backup to the other site, too. In case the DR site is destroyed….

Userlevel 5
Badge

Hi,

Yes that is possible. Depending on your setup, what you could do is copy your backups to the DR site and then run the backup to tape job there. 

Another option is that you could perform a Veeam backup to a repository at the DR site and then run the tape backup job there. However, do bear in mind that if you setup another Veeam instance, you may need additional licensing.

Yes that is possible. Depending on your setup, what you could do is copy your backups to the DR site and then run the backup to tape job there.  --- FOR THIS, i should have server and storage in dr site for veeam mgmt and repo?

Userlevel 5
Badge

OK.

you will need a server at the DR site as a Veeam tape server and connect the tape library to it.

The network connection between the primary site and the DR site has to be stable and has to have enough bandwidth to send your amount of data in your time window.

The scenario is a standard scenario, no problem.

But please keep in mind that you make a configuration backup of your VBR database and copy it to the DR site, too. With this you can setup a VBR server at the DR site and restore the backups.

 

On the other hand, you can move the VBR server to the DR site, too and keep a simple repository server at the primary site for your primary backups. Then you have complete backup environment at your DR site in a disaster. But here you should copy a configuration backup to the other site, too. In case the DR site is destroyed….

---- Thank you for this!

We have to remain the vbr server and primary storage (90tb) in Primary site for faster restoration. Then relocate the tape going to DR site.

For that design no need to have storage in the DR site (that is equivalent to 90tb)? as secondary storage, only Veeam Tape Server?

Userlevel 7
Badge +8

@JustineP, VBR server location won’t have any impact on restoration, unless the backup repository exists on the same server as the VBR, if those are separate, you can safely move VBR offsite.

 

As for the second question, you will be heavily relying on good network performance to avoid any secondary storage at the DR site, and instead writing from primary storage directly to tape.

 

Have you ever heard of tape “shoe-shining” or “back-hitching”?

If you can’t supply data to the tape drive fast enough, the tape drive will have to rewind the tape to the last point that data was written, and then continue. Odd moments of this are going to be quite likely when leveraging any kind of site-to-site connectivity, but excessive amounts are going to degrade/wear out the tape & tape drive faster and can decrease tape lifespan.

 

What LTO tape standard are you using? And what is your inter-site connectivity, is it shared bandwidth between other tasks, or purely for backup? If it’s shared, how much of that bandwidth can you reserve?

Userlevel 7
Badge +8

OK.

you will need a server at the DR site as a Veeam tape server and connect the tape library to it.

The network connection between the primary site and the DR site has to be stable and has to have enough bandwidth to send your amount of data in your time window.

The scenario is a standard scenario, no problem.

But please keep in mind that you make a configuration backup of your VBR database and copy it to the DR site, too. With this you can setup a VBR server at the DR site and restore the backups.

 

On the other hand, you can move the VBR server to the DR site, too and keep a simple repository server at the primary site for your primary backups. Then you have complete backup environment at your DR site in a disaster. But here you should copy a configuration backup to the other site, too. In case the DR site is destroyed….

---- Thank you for this!

We have to remain the vbr server and primary storage (90tb) in Primary site for faster restoration. Then relocate the tape going to DR site.

For that design no need to have storage in the DR site (that is equivalent to 90tb)? as secondary storage, only Veeam Tape Server?

Yes you would move the tape server to the DR site as noted and connect it to the VBR server.  Then the tape jobs would run as normal but already be in the DR site so no need to vault.

Userlevel 7
Badge +3

A lot depends on networking and how much data you want to replicate.


If you had a VBR server Repo, and Proxy at DR, you could have a Tape proxy and TAPE at the main site. Customers could export tapes easier.   You would be replicating data back and forth however.

 

If you had a SAN at each site dedicated to Veeam, (I don’t know how much data you have to deal with) You could have VBR at DR site (recommended for DR) A proxy at both sites and SAN at both sites.

 

Jobs backup from main site to SAN, then Veeam copy to the DR site with incremental saving time and bandwidth. You could now do Tape jobs at the main site and still be following the 3,2,1 rule while exporting tapes local.

I use a similar approach with tape a both sites. 

 

 

Userlevel 7
Badge +7

OK, many very interesting suggestions….

But to the original question:

  • yes, the tape library can be moved to the DR site - if the network connection between the sites is stable and ha enough bandwidth.
  • a server is needed at the side the tape library resides - either the VBR server or a dedicated tape server.
  • The server with the tape server service should have some storage to stage the data from tape in restore case.
  • if the VBR server is on the DR side a server at the main site is needed to manage the primary repositories - either on a dedicated storage system or on internal storage in the server.
  • configuration backups of the VBR database are needed and they have to be copied to both sides. With them you are able to recreate your VBR server in the case the site is destroyed where the VBR server resides or the VBR server itself is destroyed.

Main question is if the network connection between the sites is stable and strong enough to transport the backups in a reasonable time.

Userlevel 7
Badge +2

 

But please keep in mind that you make a configuration backup of your VBR database and copy it to the DR site, too. With this you can setup a VBR server at the DR site and restore the backups.

 

On the other hand, you can move the VBR server to the DR site, too and keep a simple repository server at the primary site for your primary backups. Then you have complete backup environment at your DR site in a disaster. But here you should copy a configuration backup to the other site, too. In case the DR site is destroyed….

 

This is always my recommendation…..keep the VBR server at the recovery site if possible because it’s easier to recover if there is a full site failure.  The VBR server could also be the tape server.  Then just have a repo/proxy server at the primary site.  At the very least, keep a small repo at the recovery site to send your configuration database too. 

Also, since you’re keeping the tapes at the recovery site, depending on what data you’re keeping in the local repo at the primary site, note that recovering from tape is going to be slower, and even slower yet over the WAN link.  Just an expectation to set up front that tape restores will take longer than a local repo restore.

Userlevel 7
Badge +3

I agree.  Either way, you will need servers at both sites,    You can use a proxy for Tape and storage, or if you really wanted configure your VBR server as a proxy/ repo or both.

 

Once again it goes to the size of your environments.   Using direct SAN I needed physical proxies, but because I bought pretty monster servers they are also my repos.  They are ALSO my tape servers currently until the last TSM servers get decommed and repurposed for Veeam. 

 

Tossing some disk in a server and adding it as a repo for staging is totally fine, but is there a reason you need it directly on the server? 

 

You could also set up a small scheduled task to copy your config as well. 

Comment