vSAN1

Hi,

I find supported configs a bit confusing for streched clusters.

Is 2+2+1 a supported streched cluster configuration?

Also as I understand 1+1+1 is supported, correct?

Thanks

Best regards!

Hello @Bylla,

Welcome to Communities.

Yes both are supported all the way up to 30 data-nodes total between both sites (+ witness).

1+1+1 can use any vSAN license level but any larger requires vSAN Enterprise license.

Bob

Hi Bylla​,

Yes 1+1+1 (2 nodes + vSAN Witness Host) is a supported configuration.

Each node is configured as a vSAN Fault Domain. It's the ROBO(Remote Office Brach office) 2 node cluster.

You can find a use case in this link with configuration example.

Also storagehub ​(the link is pointed to stretched cluster and two node configuration) could be a good resource for more deep understanding.

Please let me know if you need further assistance.

Best regards,

Hassan BEN TAHER

Thanks for the fast response @TheBobkin and @HassanBENTAHER

Do any of you have any links to share regarding the support of 2+2+1, thanks again!

Yes Bylla it is supported but in this case you will have PFTT=1(data will be distributed accross both two data sites) and SFTT=0(because in this case the minimum requirement for SFTT is to have at least 3 nodes per data with no self healing possiblity(whitch require minimum 4 host with SFTT failure methode miroring).

So you can start with is 3+3+1 as mimum requirement to have failure tolerence across sites and host failure tolerence at local site.

Concerning the self healing, in case of FTM= mirroring it require minimum 4 nodes because in case of one node failure, vSAN cluster will be able resync the data and have all your components compliant with the vSAN storage policy. In this case, even with one node failure, you will have data on host1 witch is duplicated in host2 and the witness object on host3.

regarding the link let me share with you : link1 and link 2.

Please let me know it you need further assistance.

Best regards,

Hassan BEN TAHER

If I have understood correctly, for protection against Availability Zone failure, we need a minimum 4+4+1.

So all below configuration could be right use cases for Robo.

1. 1+1+1
2. 2+2+1
3. 3+3+1 (could fall into robo and stretched) but of course with stretch it is risk.

Hello Preetam,

For the 4+4+1, it's à recommendated configuration because it will prevent from site failure and local failure and with self healing.

For the ROBO is for remonte office(as a use case) and it's only requiere only two hosts(no more).

You can double check this blog as example:

https://www.tilkens.com/2020/01/vsan-robo-project-design-config/

Best regards,

Hassan BEN TAHER

Right, but sometime we also need 2+2+1 or 3+3+1 clearly driven by sizing requirement?

Thanks for sharing the website. It is amazing. Great one.

Sure