Manual Decommissionning Procedure#

Occasionally, clusters find themselves in a broken state which renders the normal decom procedure useless due to a series of errors.

In this case, proceed with the decom of the configurations as much as possible. The steps below should be executed after removing the fe_base configuration.

Remove fe_buckets#

Connect to the FE WebUI and go to Object storage.
Search for all the buckets containing the name of the zone to be removed
Remove the S3 buckets

Check and remove all other FE resources#

In the WebUI verify that there are no resources left. If any found, delete them by hand

Clean up keycloak#

Clean up keyclock infra#

Connect to the keycloak of the infra zone using the admin account (infra-prd or infra-stg)
The password for the admin account can be found in the keycloak-credentials secret
```
kubectl -n keycloak get secret keycloak-credentials -o yaml
```
In the Identity Providers menu, select caascad operators and go to the mappers tab
Delete mappers containing the name of the target zone.
In the Clients menu, delete all clients containing the name of the target zone
In the Roles menu, delete all the roles containing the name of the target zone
In the Authentcation menu, delete the flows conaining the name of the target zone

Clean up keycloak corp#

Connect to the keycloak on the corp zone using the admin account
The password for the admin account can be found in the keycloak-credentials secret
```
kubectl -n keycloak get secret keycloak-credentials -o yaml
```
In the Roles menu, delete all the roles containing the name of the target zone

Clean up vault infra#

Connect to vault infra (infra-prd/infra-stg)
Delete the secret **secret/zones/fe/

Clean up AWS IAM#

Connect to the AWS console (prod or stg)
In the IAM service, delete all users which contain the name of the target zone