Solved

Backup Design Recommendations


Userlevel 1

Hello Veeam Gods,

Newbie here, looking for some advice on system design… I’m hoping I’m in the right place, but let me know if I’m not.

Here’s the stack:

DC1:

3x ESXi Physical Hosts

1x VCenter Physical Server

1x Win22 100TB Storage

Non-Prod Nimble System with 40TB storage

All on 25GB backbone

 

LINK: 100gb fiber connection DC1 --> DC2. Separate buildings, power, generators, WAN etc.

 

DC2:

3x ESXi Physical Hosts (same spec as DC1)

1x Veeam Physical Server with VBR + Explorers removed

1x 600TB Ubuntu hardened Immutable Storage

All on 25GB backbone

 

WAN:

2 x 2GB redundant WAN connections (1 into each site but HA across the two, redundant firewalls too)

Wasabi Buckets for Offsite S3 storage.

 

The Plan (that I’d like guidance on):

Run Production in DC1

Use CDP to DC2 for near-real-time failover

Backup Production in DC1 to the 100TB Storage in DC1 (store short term backups - sub hourly)

Copy Job to DC2 600TB Ubuntu Immutable (store short and longer term backups - sub hourly, hourly, daily, weekly, monthly)

Copy Job to Wasabi Cloud (store daily/weekly/monthly backups)

Use Nimble System for SureBackup / Virtual Labs.

 

What am I missing / What would you recommend? I’m sure I’ve not covered everything here, but it’s the high-level architecture. VMware Hosts and Network are massively oversized, so there’s really no constraints there. We’ve been running on 1GB network until 2023, so 25/100 is not a concern.

Appreciate any guidance and suggestions that the community can offer!

Thanks all!

icon

Best answer by Chris.Childerhose 6 January 2024, 16:56

View original

5 comments

Userlevel 7
Badge +10

Hi @Champion2725 welcome in this community. Reading your specs I think you've settings all needed to follow the 3-2-1 rule. Why on DC2 you haven't a vCEnter, in VCSA appliance? In case of DR you can consider to use VM balancing and HA. 

Consider to use a VM bridge with Veeam console and removing RDP access on Veeam server. 

What kind of FS uses on Win2022 server for storage?

Let me know if you want more info 

Userlevel 7
Badge +20

Based on everything you have most things seem to be covered.  I would deploy a VC in DC2 as it will make failover if required easier than just having the ESXi hosts plus you need to have a cluster to deploy the CDP filters and that will only work with a VC deployment.

Also be sure for sizing, etc. you follow the best practices guide. Ensuring the Win2022 repo is ReFS with 64k block sizing.  Follow this page - https://bp.veeam.com/vbr

Userlevel 1

@Chris.Childerhose thanks! The 6 hosts are just in one VCenter right now, just so I’m clear, are you recommending having a second VCenter head unit in DC2 and having 2 VCenters with 3 in each? 

Userlevel 7
Badge +20

@Chris.Childerhose thanks! The 6 hosts are just in one VCenter right now, just so I’m clear, are you recommending having a second VCenter head unit in DC2 and having 2 VCenters with 3 in each? 

Yes split them as it makes DR much easier cause if you have one VC and DC1 goes offline then how do you manage DC2?  That is why it would be better to have a VC in DC2 with those three hosts.

Userlevel 1

@Andanet Thank you too!! The 2022 repository is using ReFS as the file system.

RDP is on currently, we use Jumpcloud to TOTP the logins, but we’ll probably remove RDP once we’re running.

Comment