• EN
Solved

Backup Job - Best Practices

  • 11 November 2023
  • 8 comments
  • 662 views
NemanjaJanicic
Userlevel 7
Badge +1

Hello,

Since I’m going trough a lot of Veeam documentation in the past few days I found that I can always improve our Backups in some ways.

At the moment we have each VM in separate Backup Job:
 

They are all run in the similar time window: 21:00 and 22:00

In some Veeam documentation I read that its better to group them together because of the simplicity and I/O resources used.
Can somebody give me mote info about that

Description:

  • All of these VMs are from 3 different hosts.
  • 1 VM is Linux based
  • 13 VMs are Windows Server based 2016/2019

 

What are your recommendations on this

 

The largest VMs are:

  • VM Backup FS - 12 TB
  • VM Backup SVFS1002 - 3 TB
  • They are both on the same host
icon

Best answer by coolsport00 11 November 2023, 13:50

View original

8 comments

NemanjaJanicic
Userlevel 7
Badge +1

Also what should be the sizing off the VBR server itself, what are the recommendations for this?
On top of the Backup Jobs we will have Backup Copy Jobs also.
 

coolsport00
Userlevel 7
Badge +17

Hey @NemanjaJanicic ... @MicoolPaul kinda shared about your job  usage questions in a post a few mos back with regards to all VMs in 1 job vs segregated. He was testing VBR performance of pending tasks. My recommendation is for large file servers of around 3-4TB & larger, put those in their own job so as not to delay getting smaller VMs backed up. Then group all your smaller VMs in a job. When using per-VM Repo setting, you can have about 300 VMs in a job as noted in the Veeam BP Guide.

You can look there for VBR sizing as well. BTW, are you using an all-in-one VBR setup to where your Veeam server is also your Proxy & Repo? If so, look in the BP Guide on sizing for those other 2 roles as well. Whatever compute you give your VBR server, you'll have to add compute for the Proxy & Repo roles as well.

Hope that helps. 

Shane Williford - Veeam VMCA/VMCE | Veeam Legend | VUG Leader | VCP-DCV | Twitter: @coolsport00
NemanjaJanicic
Userlevel 7
Badge +1

Hey @coolsport00 ,

Repo’s will be connected via iSCSI to VMs specifically for that then connected to VBR as repositories.

We don’t have any off-host proxies.
Plan is to have off-host proxy for our largest VM.

For Proxies I think that our VBR is proxy by default and on-host proxies should be.
 

 

coolsport00
Userlevel 7
Badge +17

@NemanjaJanicic yep, that sounds like a very viable implementation design plan to go with. Just reference the User Guide for Sys Req’s , and the BP Guide for Sizing and you should be golden

Shane Williford - Veeam VMCA/VMCE | Veeam Legend | VUG Leader | VCP-DCV | Twitter: @coolsport00
Moustafa_Hindawi
Userlevel 7
Badge +6

It’s recommended to group the same os vms in one backup job for better deduplication ratio.

Moustafa Hindawi | Blogger | Nutanix Technology Champion | NCSE | NCM-MCI | VMCE | 2xVCAP-DTM & DCV | 2xVCP-NV & CMA | LinkedIn: https://www.linkedin.com/in/moustafaadelmoustafa/ | My Blog: https://cloudwhales.co.uk/author/mhindawi/
BertrandFR
Userlevel 7
Badge +8

It’s recommended to group the same os vms in one backup job for better deduplication ratio.

Hello i don’t understand your answer, this is only the case if @NemanjaJanicic is using single file backup.

Backup Chain Formats - User Guide for Microsoft Hyper-V (veeam.com)

The organisation of backup jobs can sometimes be adapted to several factors:
- Internal organisation 
- Workload (AAIP, indexing, huge servers...)

@coolsport00 I like what Shane suggested :). BTW from my understanding is the 300 limits was before V12. I played with this limit in the past 😅

You can add all your 3 hypervisors in a single jobs and exclude specifics servers and process them to a dedicated jobs. All new workload created on these hypervisors will be backuped, it could give you some flexibility.

coolsport00
Userlevel 7
Badge +17

Hi @BertrandFR - the 300 is in the BP Guide here , but yes...this was way before v12. We talked about VMs per job in the VMCA v1 Course back in I believe 2019...and this number was thown out there way back then too 😊

Shane Williford - Veeam VMCA/VMCE | Veeam Legend | VUG Leader | VCP-DCV | Twitter: @coolsport00
coolsport00
Userlevel 7
Badge +17

Hi @NemanjaJanicic - did we answer your question for you? Let us know if not.

Thanks!

Shane Williford - Veeam VMCA/VMCE | Veeam Legend | VUG Leader | VCP-DCV | Twitter: @coolsport00

Comment


Terms & Conditions