Unable to create objects and run snapshot tasks in a 2-node ROBO cluster with one of the data nodes in Maintenance Mode or not available

Products

• VMware vSAN

Issue/Introduction

After losing a host or after putting one of the data nodes in Maintenance Mode in a 2-node ROBO cluster it is not possible to create new objects or take snapshots or snapshot based backups.

Depending on the situation a "Not enough fault domains." or "Operation not allowed in current state" can be observed or the task will fail without any messages.

Environment

vSAN 7.X
vSAN 8.X

Cause

It is expected for the task to fail for an FTT=1 policy as it is not possible to satisfy the policy conditions, while one of the hosts is not taking part in the cluster during an outage or maintenance task.

Resolution

To Workaround the current policy can be edited  to include the Force Provisioning advanced Policy rule.

While the Force provisioning flag is on, the objects equipped with this policy will be provisioned regardless of the fact that the current policy is not satisfiable using the resources currently available in this cluster. vSAN will try to bring the object back to compliance when the missing resource is restored.

***PLEASE NOTE*** It is strongly recommended  to return this setting back to off as soon as the task is completed or at latest as soon as the cluster is restored back to healthy state. It might otherwise lead to unwanted behaviours and additional issues.

To turn the setting ON please follow the below steps:

1. Please identify what policy is to be adjusted (if multiple or all VM's in the cluster are to be included there might be multiple policies applicable).
2. Navigate the vSphere Client "Hamburger Menu"->VM Storage Policies.
3. Select the Policy to be changed (if there are multiple they need to be done one at a time) and click the EDIT button above.



4. please click NEXT to get to step 2 in the "Edit VM Storage Policy" wizard and open the "Advanced Policy Rules".
5. Click on the Slider next to the Force Provisioning option to set it to ON (it will turn green) then click NEXT all the way to FINISH.

6. At the end of the wizard you will be asked to reapply the policy and please choose the Now option, and proceed to the next one , if applicable.

Additional Information

If the problem lays with the witness host instead of the data nodes please see the below article:

"There are currently x usable fault domains. The operation requires 1 more usable fault domains." while taking virtual machine snapshot