What happens when a local site does not respect SFTT anymore ?

[hiluvatar]

Hi guys,

I have a question on the streched cluster topic: What happens when a local site does not respect SFTT anymore due to multiple disk/server loss? Let's say my SFTT=1 and lose 2 servers at the same time (or two disks).
A/ The entire site is marked as failed and all VM are restarted on the remaining site
B/ Only affected components are marked as failed and the cluster use remote resources to restart VM (or rebuild data)

The documentation is not clear on the topic.

[TheBobkin]

Hello hiluvatar​,

Welcome to Communities.

If you have SFTT=1 on a stretched cluster then either site can withstand losing a component locally and then this site is technically FTT=0 until it gets access to the data again or rebuilds it.

If you lose 2 components in one site then some Objects will become inaccessible from the data on this site, but provided the other site is still FTT=1 (or even FTT=0) the Objects will be accessible from the data there. If the data is still accessible, restarting VMs shouldn't be necessary.

Bob

[hiluvatar]

Hello,

thanks for your answer.

This is where I don't understand the documentation :

Source : https://storagehub.vmware.com/export_to_pdf/vsan-stretched-cluster-2-node-guide

+ in the same document it's said : "If a site has more failures than the local protection policy will allow, then the site is considered failed."

[hiluvatar]

Hello guys,

do you have an update on the topic?

Thanks for your clarification.

[depping · Accepted solution]

vSAN never sees a site as fully failed unless ALL hosts in a site has failed. The explanation above is a better explanation, I have requested the documentation to be updated as it is fairly complex to read.

[hiluvatar]

Ok thanks a lot Duncan, that's what I thought.