Error trying to apply disk stripes per object at 3

[MichaelGi]

I'm getting the following error message.  It's a 4 host cluster with 7 disks per host. 2 SSD's and 7 mechanical.  Why is the error saying there are only 3 disks total in the cluster?

Insufficient disk space on datastore ''.

The policy requires 2 replicas with 3 disks(minimum version 2) each Found a total of 3 disks in the cluster. 0 disks decommisioning right now.

Storage policy change failure: There is not enough space on the file system for the selected operation.

[CHogan]

It sounds like the cluster is partitioned or not formed properly.

A few questions to get started.

1. Which version of VSAN?

2. If 6.0 or later, have you health check? Does it report any issues?

3. Can you deploy any VMs? For example, if this is a POC or lab, can you run the proactive test to create VMs. Does that pass?

4. What is the actual policy that you are trying to deploy with? What is the number of failures to tolerate? what is the stripe width?

5. If using the default policy, can you check that the default policy is set to?

6. how are the physical devices presented to the host? Are they pass-thru or in a RAID-0 config? If RAID-0, are each individual disk in its own RAID-0 volume?

[MichaelGi]

Thank you very much for responding to my issue.

I'm using VSAN version 6.1.

The health check passes.

There are around 20 vm's running in production using the default policy of FTT = 1 and Stripe Width = 1

One of the vm's is using a policy of FTT = 1 and Strip Width = 3 for better write performance.

I can't change any of the other VM's to this policy because I get the error message mentioned previously.

The devices are presented as each disk as RAID-0.  Each disk has it's own RAID-0.

There are 2 disks groups per hosts with 1 ssd and either 3 or 4 mechanical disks per group.

I've tried creating new VM's with this policy and I get the same error.

There's 11 TB free in the cluster.

All proactive tests pass.

Update:

I found that the vsantraces folder was full on 3 of the hosts.  There were several vsanObserver files in it.  I cleaned these folders up and was able to set the new policy.