Is the migration hanging?

[johnpangalos]

Hey,

We have just inherited a weblate server running version 3.9.1 (docker version 3.9.1-1), and we just moved it from a server to kube using the helm chart and setting the image tag property to 3.9.1-1.

This seems to work, but now we're trying to upgrade weblate to version 4.0.4 (docker version 4.0.4-5) and we are running into issues. We get kube to start the docker container and it seems to connect to the postgres instance but after that it seems to hang. Here are the logs we get:

Starting...
weblate
Postgres is up
System check identified some issues:

INFOS:
?: (weblate.I031) New Weblate version is available, please upgrade to 5.5.5.
        HINT: https://docs.weblate.org/en/weblate-4.0.4/admin/upgrade.html
Starting database migration...

We have had the same results with version 3.11.3 as well. Is this normal behavior? I was expecting more logs to as what the migration is doing.

Here is our helm chart:

# Helm chart reference:
# https://artifacthub.io/packages/helm/weblate/weblate
emailHost: <Redacted>
emailPort: 587
emailTLS: true
emailSSL: false
emailUser: <Redacted>

defaultFromEmail: <Redacted>
serverEmail:  <Redacted>
adminEmail:  <Redacted>

persistence:
  size: 100Gi

image:
  tag: 3.9.1-3

service:
  type: NodePort

ingress:
  enabled: true
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-production
    kubernetes.io/ingress.allow-http: "false"
  hosts:
    - host: <Redacted>
      paths:
        - path: /
          pathType: Prefix
  tls:
    - hosts:
        - <Redacted>
      secretName: <Redacted>

extraConfig:
  WEBLATE_LOGLEVEL: DEBUG

externalSecretName: external-secret
existingSecret: existing-secret

postgresql: 
  enabled: false
  postgresqlHost: <Redacted>
  auth:
    existingSecret: existing-secret
    database: weblate
    enablePostgresUser: true
  service:
    ports:
      postgresql: 5432

redis:
  auth:
    existingSecret: existing-secret

resources:
  limits:
    cpu: 4
    memory: 10Gi
  requests:
    cpu: 2
    memory: 8Gi

livenessProbe:
  initialDelaySeconds: 72000
  periodSeconds: 30
  failureThreshold: 10

[nijel]

Some migrations are expected to take long, always consult upgrading instructions for that.

[johnpangalos]

Thanks for the quick response, so the output was normal? There is no indication of progress? Just checking to see if this is intended behavior.

I did check the docs and it did say that it would take a long time for the specific upgrade we were attempting.

[johnpangalos]

We checked the database and it does seem to be performing some actions even though they aren't being logged. Thanks for the help!

[nijel]

The Docker logging output is buffered, so you only see the output once the migration step is completed.