AWX migration issue at task Stream backup from pg_dump to the new postgresql container

[battlestargalactica]

Hello - I am having some problems now trying to get past this task. Whats different in my migration - i have an existing awx instance with a managed postgres-db pod in an eks cluster.

I have a brand new build for an external RDS postgresql instance.

I created the old secret and the new secret for the external rds db

(awx-prod-old-postgres-configuration & awx-prod01-postgres-configuration)

I applied the following with kustomize cmd:

 kustomize build . | kubectl apply -f -

here is my kustomization file:

apiVersion: kustomize.config.k8s.io/v1beta1

kind: Kustomization

resources:

  # Find the latest tag here: https://github.com/ansible/awx-operator/releases

  - github.com/ansible/awx-operator/config/default?ref=1.1.0

  - awx-prod.yml

# Set the image tags to match the git version from above

images:

  - name: quay.io/ansible/awx-operator

    newTag: 1.1.0

# Specify a custom namespace in which to install AWX

namespace: awx

And here is my awx-prod.yml deployment

apiVersion: awx.ansible.com/v1beta1

kind: AWX

metadata:

  name: awx-prod

  namespace: awx

spec:

  service_type: NodePort

  nodeport_port: 30080

  ingress_type: none

  service_labels: |

    environment: production

    team: *******

  hostname:  *******

  old_postgres_configuration_secret: awx-prod-old-postgres-configuration

  postgres_configuration_secret: awx-prod01-postgres-configuration

  secret_key_secret: awx-prod-secret-key

I watched the scale down occur and now it seems I am stuck at this part it seems - it has been sitting here for couple of hours:

--------------------------- Ansible Task StdOut -------------------------------

TASK [installer : Stream backup from pg_dump to the new postgresql container] ***
task path: /opt/ansible/roles/installer/tasks/migrate_data.yml:57

-------------------------------------------------------------------------------
{"level":"info","ts":1671039027.6071534,"logger":"proxy","msg":"Read object from cache","resource":{"IsResourceRequest":true,"Path":"/api/v1/namespaces/awx/pods/awx-prod-postgres-13-0","Verb":"get","APIPrefix":"api","APIGroup":"","APIVersion":"v1","Namespace":"awx","Resource":"pods","Subresource":"","Name":"awx-prod-postgres-13-0","Parts":["pods","awx-prod-postgres-13-0"]}}

I dont see anything happening on the RDS side - there are no active sessions - does my RDS need to be set to manage by the operator? here is what I have configured for the database:

---

apiVersion: v1

kind: Secret

metadata:

  name: awx-prod01-postgres-configuration

  namespace: awx

stringData:

  host: "anrdsdatabasehostname.us-region.rds.amazonaws.com"

  port: "5432"

  database: "awx_prod_db"

  username: "postgres_admin"

  password: password_noquotes

  sslmode: "prefer"

  type: "unmanaged"

type: Opaque

[battlestargalactica]

are we able to have the operator migrate outside of the k8s to the rds database?

[battlestargalactica]

I ended up going the pgdump | pgrestore method - had to make sure the secret for the RDS was created beforehand and re-applied the AWX object into the cluster, also needed the external rds created with the same role that the k8s postgresdb was using called "awx", otherwise there are errors thrown for ownership on the tables during the pgrestore to the external db. 

[AWX Project]

glad you figured it out! thanks for sharing what you did to make it work

AWX Team