Vsan storage policy

[Timur1991]

Hello. Tell me please.
I have a vsan cluster of 4 servers and 1 witness.
The FTT -1 RAID 1 - Mirroring storage policy is set.
For some reason, some machines in the Monitor - Physical disk placement section, some machines are listed as VM hard disk - RAID 1, and some are RAID 1 and below it is broken down by servers and says RAID 0. I want to know if my machines are protected by the FTT - 1 policy?

Screenshot attached

[depping]

Yes, you have RAID-1 with RAID-0 for that VM as you see. You probably have a stretched cluster, with 2 hosts in the preferred fault domain and two hosts in the secondary fault domain. Judging by the disks, you probably configured to protect against a site failure, and then you also have a stripe width set, or the disks are very large and as such need to be spread across hosts.

But yes, you are protected with RAID-1 against a fault domain / site failure.

[Timur1991]

Thanks for the reply. Yes, you said everything correctly. This is a sprawling cluster with two hosts in the preferred and secondary fault domain and one witness host. The Vlan default storage policy is used. What worries me is that it says RAID 0. This happens only with cars that weigh a lot

[depping]

If a VM Disk is larger than 256GB than vSAN will split it up in multiple components. Those components can be stored on the same host, as the screenshot shows, or can be stored on a different host (which the screenshot also shows).

So there's nothing to worry about here. You are protected as specified in policy.

[Timur1991]

Thanks, friend

[Timur1991]

I have a question that only you can answer. In such a cluster, if a disk group fails on the preferred domain, but the host is still alive, does the machine continue to work from the surviving replica located on the secondary site? or restarts on the secondary domain? + the question is if this is the case, does the rebuild of the machine object take place?
https://dropmefiles.com/nxdd0
https://dropmefiles.com/9zP6n
Screenshots 

[Timur1991]

Just worried that it says RAID 0... the system is already in production. So I think if RAID 0 can take action before the drive fails...

[depping]

your policy makes no sense to me. your policy doesn't state that the data is supposed to be stretched. anyway, the data that you showed is stretched based on what the other screenshot shows, and that is what vSAN does by default in this case.

Please note, YOU DO HAVE RAID-1, It says so at the top of the screenshot. But then you have RAID-0. So ALL data is replicated.

Nevertheless, I think you need to ask yourself a few things:

• What am I trying to achieve?
  • Do I need VMs which are protected across locations?
    • If yes: Select the policy option that says "site disaster tolerance" and configure it accordingly
    • If no: Select the policy option that says "site disaster tolerance" and configure it accordingly
  • Apply policy to VMs
  • TEST!
    • What do you mean test? Well with a stretched cluster, simply test failure scenarios, kill a host and see what happens!

Anyway, yes if a diskgroup fails then the VM will remain running from the other component, and if there's sufficient diskspace in the fault domain where the host/diskgroup failed then the component will be resynced.

We have plenty of documentation on this topic:

https://core.vmware.com/

[Timur1991]

I now have PFTT = 1 and SFTT = 0, right?
These are two different sites, geographically separated, and connected via a 10G vsan network

[depping]

Yes that is what you have.

[depping]

I wrote about it a while back by the way: https://www.yellow-bricks.com/2020/12/14/vms-which-are-not-stretched-in-a-stretched-cluster-which-po...

[Timur1991 · Accepted solution]

I did a test, turned off the host the machine is running on, restarted the machine with HA on the failing host.
It is important for me that in case of a disk crash, the machine does not restart on another host ... but you answered this question. You said that if the diskgroup crashes, the machine will continue to run on the original host, not restart on another host

[Timur1991]

You just wrote here - Then there's “None - Standard Cluster”, what happens in this case? When you use “None – Standard Cluster” with “RAID-1”, what is going to happen is that the VM is configured with FTT=1 and RAID-1, but in a stretched cluster “FTT” does not exist, and FTT will automatically become PFTT. This means that the VM is going to be mirrored locations across, and you will have SFTT=0, which means no resiliency locally. It is the same as “Dual Site Mirroring”+”No Data Redundancy”!

But this is also redundancy, the fact that data is replicated between sites

[Timur1991]

Failure of the host normally fulfills without problems. The question is what happens when a disk or disk group fails in a stretched cluster with a policy of PFTT=1 SFTT=0. For example, a disk failed on the primary site, there is a copy on the secondary site, what happens in this case? the machine continues to work from a copy of the secondary site, or what is happening? I can't run a test because it's a production. All hope is on you. No one can advise you better than you...

[depping]

The VM will continue to run. The UI labels are a bit unfortunate, but you have indeed data redundancy so the VM will work, it is just that the "local copy" will be unavailable at that point in time and IO will need to go across the intersite link.

[Timur1991]

Thank you very much ... sorry for asking a lot of questions. Only you could decide

[depping]

No worries.