Note: This probably only happens on non-etcd datastores and isn't catched by the integration tests.

Reproduction

• Create a Postgres-backed datastore and create a TCP on it
• Create a second Postgres-backed datastore
• Migrate the TCP onto the new datastore
• Try to edit a resource in the child cluster

Behaviour

It should work.

Current Behaviour

The TCP is still in read-only mode, because it is still connected to the old (and read-only) datastore.
After another (a bit delayed) reboot of all TCP pods, it starts working fine.

Feel free to reach out, if you have any further questions!