Jump to content

Recommended Posts

Posted

Afternoon,

 

Fail over cluster if Cluster managed voting is Enabled do you still need quorum witness?

 

The reason i ask is because we set up a 3 node cluster but still get the "The cluster is not configured with a quorum witness. As a best practice, configure a quorum witness to help achieve the highest availability of the cluster." on validation even though Cluster managed voting: Enabled

 

if a witness is needed for 3 nodes how is that different from 2 nodes with a share witness?

This will be used for SQL.Thoughts?

This will be used for SQL.

Posted

It's not "needed" as per the warning, it's just best reliability, with dynamic quorum management you can go down to a single node and still have a proper vote. Otherwise it'd automatically be split.

 

In basic terms (ok yes if you query it a lot this isn't entirely accurate), but you need 50% of hosts up without witness. So in a 3 node, you'd need 2 hosts (without witness). With witness you can need 1+1witness, as the 2/4 combo etc

 

Steve

Posted
It's not "needed" as per the warning, it's just best reliability, with dynamic quorum management you can go down to a single node and still have a proper vote. Otherwise it'd automatically be split.

 

In basic terms (ok yes if you query it a lot this isn't entirely accurate), but you need 50% of hosts up without witness. So in a 3 node, you'd need 2 hosts (without witness). With witness you can need 1+1witness, as the 2/4 combo etc

 

Steve

So i keep thinking that a node votes in the same way that witness does. But starting to think that is not the case.

 

For example all nodes votes are special, but a witness's vote is extra special?

  • 2 weeks later...
Posted

So here is the MS guide that gives the absolute detail on this rather than me writing it out:

https://docs.microsoft.com/en-us/azure-stack/hci/concepts/quorum

 

Note the statement:

Cluster quorum recommendations

If you have two nodes, a witness is required.

If you have three or four nodes, witness is strongly recommended.

 

Strongly recommended in a 3 node cluster further down the doc outlines what scenarios the various cluster configurations can operate under and the liklihood of the cluster being able to continue in those failure scenarios.

Posted

I'd still add it in as per best practise. 200-400GB partition on the SAN is perfectly adequate.

 

Presenting it as a cluster storage also means the owner node can switch and move around like your regular pools

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...