Skip to main content
Solved

Backup direct from Primary Storage Snapshot


Forum|alt.badge.img

hi Veeam community 

I want to buy new storage for the repository and I got acquainted with the option “Backup direct from Primary Storage Snapshot”
My question is where is the best way to use this option?For what server and services it is better to use?It also has disadvantages?And finally, if you have the experience of using this feature, I would appreciate it if you could share and what kind of storage you used.

thanks

Best answer by coolsport00

Hi @miriam1989 -

You use this feature for Backup Jobs. It is best to use for Jobs which contain highly transactional VMs such as DBs because Veeam first snaps the VMs which run on a given datastore, which will then snap the datastore, to back up a Job's VMs. The advantage of this is the Snapshot taken is real quick vs the Snapshot taken using another backup methods (hotadd or directsan) can be retained a bit long. 

Another advantage of using BFSS is the ability to use Veeam to orchestrate your SAN snapshots and even replicate them to a partner array. This helps you achieve the 3-2-1-1-0 rule. 

One disadvantage of using BFSS is sometimes the initial processing phase of the job can take 3-5mins before backups of the VMs begin. 

I do use this feature myself, and have for several years. I use Nimble storage, now called Alletra, as my backend storage for this. 

Hope this helps. 

View original
Did this topic help you find an answer to your question?

11 comments

wesmrt
Forum|alt.badge.img+4
  • Veeam MVP
  • 226 comments
  • March 25, 2024

To use this option you must have a Storage that supports Snapshots and it’s supported by Veeam.
The biggest advatange is the short time the snapshot lives on vSphere side during the backup. We explain better in the User Guide: https://helpcenter.veeam.com/docs/backup/storage/vm_processing.html?ver=120

 

Also, this KB shows the Storages that supports Snapshot integration with VBR: https://www.veeam.com/kb4206


coolsport00
Forum|alt.badge.img+20
  • Veeam Legend
  • 4145 comments
  • Answer
  • March 25, 2024

Hi @miriam1989 -

You use this feature for Backup Jobs. It is best to use for Jobs which contain highly transactional VMs such as DBs because Veeam first snaps the VMs which run on a given datastore, which will then snap the datastore, to back up a Job's VMs. The advantage of this is the Snapshot taken is real quick vs the Snapshot taken using another backup methods (hotadd or directsan) can be retained a bit long. 

Another advantage of using BFSS is the ability to use Veeam to orchestrate your SAN snapshots and even replicate them to a partner array. This helps you achieve the 3-2-1-1-0 rule. 

One disadvantage of using BFSS is sometimes the initial processing phase of the job can take 3-5mins before backups of the VMs begin. 

I do use this feature myself, and have for several years. I use Nimble storage, now called Alletra, as my backend storage for this. 

Hope this helps. 


Chris.Childerhose
Forum|alt.badge.img+21

Check out the links that Wesley posted as there is good information there.

When I did some testing, we were using both the Hitachi Plugin as we are a Hitachi Partner and also, I was able to thoroughly test the Pure Storage plugin with the iSCSI implementation.  Worked extremely well with backups and used the storage snapshots from the array to make things much faster.  Great storage for sure.


Forum|alt.badge.img
  • Author
  • Comes here often
  • 23 comments
  • March 25, 2024
coolsport00 wrote:

Hi @miriam1989 -

You use this feature for Backup Jobs. It is best to use for Jobs which contain highly transactional VMs such as DBs because Veeam first snaps the VMs which run on a given datastore, which will then snap the datastore, to back up a Job's VMs. The advantage of this is the Snapshot taken is real quick vs the Snapshot taken using another backup methods (hotadd or directsan) can be retained a bit long. 

Another advantage of using BFSS is the ability to use Veeam to orchestrate your SAN snapshots and even replicate them to a partner array. This helps you achieve the 3-2-1-1-0 rule. 

One disadvantage of using BFSS is sometimes the initial processing phase of the job can take 3-5mins before backups of the VMs begin. 

I do use this feature myself, and have for several years. I use Nimble storage, now called Alletra, as my backend storage for this. 

Hope this helps. 

Thank you for sharing your experience, it is very helpful for me

 


coolsport00
Forum|alt.badge.img+20
  • Veeam Legend
  • 4145 comments
  • March 25, 2024

HI @miriam1989 ...no problem at all. Glad it helped you 😊


Forum|alt.badge.img
  • Author
  • Comes here often
  • 23 comments
  • March 25, 2024
Chris.Childerhose wrote:

Check out the links that Wesley posted as there is good information there.

When I did some testing, we were using both the Hitachi Plugin as we are a Hitachi Partner and also, I was able to thoroughly test the Pure Storage plugin with the iSCSI implementation.  Worked extremely well with backups and used the storage snapshots from the array to make things much faster.  Great storage for sure.

It was great, thank you, I will also check about hitachi, I am currently researching two brands, Hpe and Datacore.

 


Chris.Childerhose
Forum|alt.badge.img+21
miriam1989 wrote:
Chris.Childerhose wrote:

Check out the links that Wesley posted as there is good information there.

When I did some testing, we were using both the Hitachi Plugin as we are a Hitachi Partner and also, I was able to thoroughly test the Pure Storage plugin with the iSCSI implementation.  Worked extremely well with backups and used the storage snapshots from the array to make things much faster.  Great storage for sure.

It was great, thank you, I will also check about hitachi, I am currently researching two brands, Hpe and Datacore.

 

If you do look into Hitachi keep this in mind - you need double the amount of storage for it to work properly for the snapshots, etc.  I found this out while testing and investigating.


Forum|alt.badge.img
  • Author
  • Comes here often
  • 23 comments
  • March 26, 2024
coolsport00 wrote:

Hi @miriam1989 -

You use this feature for Backup Jobs. It is best to use for Jobs which contain highly transactional VMs such as DBs because Veeam first snaps the VMs which run on a given datastore, which will then snap the datastore, to back up a Job's VMs. The advantage of this is the Snapshot taken is real quick vs the Snapshot taken using another backup methods (hotadd or directsan) can be retained a bit long. 

Another advantage of using BFSS is the ability to use Veeam to orchestrate your SAN snapshots and even replicate them to a partner array. This helps you achieve the 3-2-1-1-0 rule. 

One disadvantage of using BFSS is sometimes the initial processing phase of the job can take 3-5mins before backups of the VMs begin. 

I do use this feature myself, and have for several years. I use Nimble storage, now called Alletra, as my backend storage for this. 

Hope this helps. 

 

A question about using this method for sql database Is this method suitable for transaction logs?


Chris.Childerhose
Forum|alt.badge.img+21
miriam1989 wrote:
coolsport00 wrote:

Hi @miriam1989 -

You use this feature for Backup Jobs. It is best to use for Jobs which contain highly transactional VMs such as DBs because Veeam first snaps the VMs which run on a given datastore, which will then snap the datastore, to back up a Job's VMs. The advantage of this is the Snapshot taken is real quick vs the Snapshot taken using another backup methods (hotadd or directsan) can be retained a bit long. 

Another advantage of using BFSS is the ability to use Veeam to orchestrate your SAN snapshots and even replicate them to a partner array. This helps you achieve the 3-2-1-1-0 rule. 

One disadvantage of using BFSS is sometimes the initial processing phase of the job can take 3-5mins before backups of the VMs begin. 

I do use this feature myself, and have for several years. I use Nimble storage, now called Alletra, as my backend storage for this. 

Hope this helps. 

 

A question about using this method for sql database Is this method suitable for transaction logs?

Not sure I would look at storage snapshots for transaction logs with SQL.  You may want to look at another backup method like using the agent or CDP.


coolsport00
Forum|alt.badge.img+20
  • Veeam Legend
  • 4145 comments
  • March 27, 2024
miriam1989 wrote:
coolsport00 wrote:

Hi @miriam1989 -

You use this feature for Backup Jobs. It is best to use for Jobs which contain highly transactional VMs such as DBs because Veeam first snaps the VMs which run on a given datastore, which will then snap the datastore, to back up a Job's VMs. The advantage of this is the Snapshot taken is real quick vs the Snapshot taken using another backup methods (hotadd or directsan) can be retained a bit long. 

Another advantage of using BFSS is the ability to use Veeam to orchestrate your SAN snapshots and even replicate them to a partner array. This helps you achieve the 3-2-1-1-0 rule. 

One disadvantage of using BFSS is sometimes the initial processing phase of the job can take 3-5mins before backups of the VMs begin. 

I do use this feature myself, and have for several years. I use Nimble storage, now called Alletra, as my backend storage for this. 

Hope this helps. 

 

A question about using this method for sql database Is this method suitable for transaction logs?

Hi @miriam1989 - somehow I missed this other question. Sure...you can use this method for Transaction log processing. Any backup method can be used for this → nbd, hotadd, directstorage, or bfss. Transaction log processing is configured in the Guest Processing > AAIP section, Advanced button of your Jobs. I don’t personally backup any of my DB VMs using AAIP, but yes...it can be configured.

Hope this helps.


Scott
Forum|alt.badge.img+9
  • Veeam Legend
  • 1003 comments
  • March 28, 2024

Almost all jobs will start after a few minutes as CoolSport said.  I use BFSS for all my jobs however. 

 

It removes the stun if you have large VM’s. I have very large servers and databases that are sensitive to VM stun. 

 

It has VMware take a snapshot, the SAN will take a snapshot, then the VM snapshot is removed immediately.  Veeam continues to get the data from the storage snapshot not effecting production. 

 

You can also do storage snapshot only jobs, or tell Veeam to leave the storage snapshots behind for a set amount of time after completion.  These can be set to immutable at SAN level as well.

 

I have storage snapshots on almost all my SAN volumes with Veeam allowing for very frequent rolling restore points, with backups going to other storage. If you have a second SAN at an additional location, you can run a storage snapshot job to a global mirror or secondary site as well using a secondary job which is incredible.

 

One thing to take note of, if you have small VM’s with minimal change, sometimes using a different transport mode will complete faster.  It’s good to test multiple methods and see what works best for what workload.  Veeam has done a great job of improving this over the years and not leaving snapshots behind to manually delete.

 

Look at the Storage limitations for the following products here. My IBM equipment has worked fantastic. 

https://helpcenter.veeam.com/docs/backup/storage/storage_specific_reqs.html?ver=120