Once the the node is restored from a back-up state, as it may not have the latest Federation state, the recommendation is to evict (force leave) the SDDC Manager instance from the Federation and join the instance back if required.
This can be achieved using the Multi-Site Evict API.
Note: Evict should be executed from SDDC Manager which has role as 'Controller' in the Federation.
Below are the steps to execute the Evict API.
- Login to SDDC Manager (which was assigned as Controller in the Federation) as user 'vcf'.
- Execute the below API from get an Access Token to be used for later steps:
curl 'https://<SDDC-Manager-FQDN>/v1/tokens' -i -X POST \
-H 'Content-Type: application/json' \
-H 'Accept: application/json' \
-d '{
"username" : <admin username>,
"password" : <admin password>
}'
- Execute the Evict API from SDDC Manager using the acquired Access Token.
curl 'https://<Controller-SDDC-Manager-FQDN>/v1/sddc-federation/members/<<Evicted-SDDC-Manager-FQDN>>?force=true' -i -X DELETE \
-H 'Authorization: Bearer <access token>'
- Navigate to the Multi-Site Dashoard and confirm the evicted instance is no longer visible.
Note: For information on the VMware Cloud Foundation API, see the API documentation guide .
- Add the restored instance back into the Federation from the Multi-Site Dashboard.
- Confirm that all the existing instances in the Multi-Site Dashboard should be now healthy (Green).
- Re-try any failed Join & Leave operations.