2đź‘Ť
Solve.
Thank you everyone for your help. The problem is I’m not have enough RAM in order to run the migration script, so I upgrade the VPS to 8GB RAM and then it’s work.
Because the migration script have some memory leak so pods need at least 4GB free memory to run. See this comment on Github
5đź‘Ť
you must migrate schema’s manually,
run the following command to migrate your schema:
kubectl exec -it <your-pod-name> sentry upgrade
0đź‘Ť
Faced a similar issue. In our case, we have enabled ingress (Nginx). And this setting conflict with service.type: LoadBalancer, just change it to ClusterIP and do the clear setup again.
TLDR – DB init starts on “post-install” event (all deployed components is in a ready state) but in case of LoadBallancer, ingress and --wait combination the web Service must have assigned IP (https://github.com/helm/helm/blob/master/docs/using_helm.md#helpful-options-for-installupgraderollback). For some reason it wasn’t true and web service always returns false. After change the service setting from LoadBalancer to ClusterIP the issue is gone.