bug: schema out of sync on workers after deleting a node from the schema

[wvandeun]

Component

API Server / GraphQL

Infrahub version

1.0.1

Current Behavior

The schema between different workers goes out of sync, when you delete an existing schema node from the schema.

Expected Behavior

All the workers load the new schema and the schema is in sync between them.

Steps to Reproduce

• load an instance of infrahub
• load this schema

---
version: "1.0"
nodes:
  - name: Device
    namespace: Network
    human_friendly_id: ['hostname__value']
    attributes:
      - name: hostname
        kind: Text
        unique: true
      - name: model
        kind: Text
  - name: Interface
    namespace: Network
    attributes:
      - name: name
        kind: Text
      - name: description
        kind: Text
        optional: true

• execute this GraphQL query, the schema has should be in sync, take note of the hash of one of the nodes

query {
  InfrahubStatus {
    summary {
      schema_hash_synced
    }
    workers {
      edges {
        node {
          schema_hash
        }
      }
    }
  }  
}

• load this schema, the NetworkInterface should be deleted from the schema

---
version: "1.0"
nodes:
  - name: Device
    namespace: Network
    human_friendly_id: ['hostname__value']
    attributes:
      - name: hostname
        kind: Text
        unique: true
      - name: model
        kind: Text
  - name: Interface
    namespace: Network
    state: absent
    attributes:
      - name: name
        kind: Text
      - name: description
        kind: Text
        optional: true

• execute the following Graphql query

```graphql
query {
  InfrahubStatus {
    summary {
      schema_hash_synced
    }
    workers {
      edges {
        node {
          schema_hash
        }
      }
    }
  }  
}

The workers are now not in sync anymore. Only 1 worker will have a new schema hash, the other workers still have the old hash from the previous step.

Additional Information

No response

[BaptisteGi]

Same issue but different way to trigger it:

• load schema to main
• create a branch
• load schema to delete the NetworkInterface on the new branch
• merge to main

=> We still have the deleted node on main (same schema out of sync)
=> A reboot of infrahub does put all workers on the same page

https://discord.com/channels/1212332642801025064/1306016379102560328