Helm / Tiller Terminates Hook Job too quick before it's complete
Closed this issue · 1 comments
t83714 commented
Helm / Tiller Terminates Hook Job too quick before it's complete
It only happened when #7 has happened (i.e. redeploy to an interrupted release).
Seem this issue with helm since 2.15:
To reproduce:
- Deploy using
terraform apply - Once you see error as described in #7 , re-deploy using
terraform apply - Go to cloud console and
kubectl -n magda get pods -w - you will see migrators are created quite late and are terminating before complete (in a few seconds)
t83714 commented
Caused by the newer version terraform helm provider that had been released a few days ago.
Fixed by this commit: lock helm provider to 0.10.2 to avoid the bug of helm v2 after 2.15