[stable/postgresql] data directory "/bitnami/postgresql/data" has invalid permissions

[acim]

Describe the bug
After upgrade from 8.3.4 to 8.4.0 pod is crashing with message:

postgresql 12:46:44.18 INFO  ==> ** Starting PostgreSQL **
2020-02-21 12:46:44.204 GMT [1] FATAL:  data directory "/bitnami/postgresql/data" has invalid permissions
2020-02-21 12:46:44.204 GMT [1] DETAIL:  Permissions should be u=rwx (0700) or u=rwx,g=rx (0750).

Version of Helm and Kubernetes:

Helm 3.1.1
Kubernetes 1.17.3

Which chart:

stable/postgresql 8.4.0

What happened:

Crash.

What you expected to happen:

Run.

How to reproduce it (as minimally and precisely as possible):

Just upgrade from 8.3.4 to 8.4.0.

Anything else we need to know:

Init container init-chmod-data seems to be removed and it seems it is still needed.

[jim-barber-he]

I just hit this myself.
The default for volumePermissions.enabled changed from true to false.
Setting it to true should get things going again.

[jim-barber-he]

It was broken by this change I think.
#20910

[javsalgar]

Hi,

Good to know that it was just enabling the volumePermissions container. As stated in the PR, this is not compatible with all platforms that's why we're setting it to false by default. Do not hesitate to open a new ticket if you find any other issue.

[acim]

It worked, but I had to uninstall the chart first and then install again. It seems this init is not executed on each start (I didn't look at the code).

[jim-barber-he]

FYI, for us, we have a pretty vanilla install of Kubernetes provisioned into an AWS environment using Kops.
I would have thought that would be a reasonably common scenario, so not sure why it broke for us.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Any further update will cause the issue/pull request to no longer be considered stale. Thank you for your contributions.

[stale]

This issue is being automatically closed due to inactivity.

[kuisathaverat]

it happens from time to time when we run helm upgrade without changes, when it happens we uninstall the helm chart and install it again, it is an old issue thing recurrent and unresolved bitnami/charts#1649 https://github.com/bitnami/bitnami-docker-postgresql/issues/176 #13651 #14538 bitnami/charts#1210

[T1loc]

Same here. With this values :

---
image:
  tag: 10.6.0
persistence:
  size: 8Gi
postgresqlPassword: {{ secrets.postgresql.POSTGRES_PASSWORD }}
volumePermissions:
  enabled: true

Running :

kubectl exec -it postgresql-postgresql-0 -- chmod 700 /bitnami/postgresql -R 

As a temporary fix...