Cannot write into "config" directory! Error when enabling persistence

[SeveHo]

Hello,

when enabling persistence I get the following error:

Configuring Redis as session handler
Initializing nextcloud 22.2.3.0 ...
Initializing finished
New nextcloud instance
Installing with SQLite database
starting nextcloud installation
Cannot write into "config" directory!
This can usually be fixed by giving the webserver write access to the config directory

The values when using the pvc managed by the chart are looking like this:

nextcloud:
    configs:
      custom.config.php: |-
        <?php
        $CONFIG = array (
          'overwriteprotocol' => 'https',
          'overwrite.cli.url' => 'https://<REDACTED>.<REDACTED>.com',
          'loglevel' => '2'
        );

persistence:
  enabled: true
  annotations:
    helm.sh/resource-policy: keep
  storageClass: "csi-cinder-high-speed"

Did anyone face the same issue or sees what I am doing wrong?

[vasanthaganeshk]

Have you tried with deleting and rerunning the helm install?

[SeveHo]

Yes. A lot of times.

[android10]

Facing the same problem here. Folder permission:

drwxr-xr-x fernando fernando 4.0 KB Tue Feb 22 00:06:07 2022  .
drwxrwxrwx fernando fernando 4.0 KB Mon Feb 21 23:59:26 2022  ..
drwxr-xr-x root     root     4.0 KB Tue Feb 22 00:06:07 2022  config
drwxr-xr-x root     root     4.0 KB Tue Feb 22 00:06:07 2022  custom_apps
drwxr-xr-x root     root     4.0 KB Tue Feb 22 00:06:07 2022  data
drwxr-xr-x http     root     4.0 KB Tue Feb 22 00:14:14 2022  html
drwxr-xr-x root     root     4.0 KB Tue Feb 22 00:06:08 2022  root
drwxr-xr-x root     root     4.0 KB Tue Feb 22 00:06:07 2022  themes
drwxr-xr-x root     root     4.0 KB Tue Feb 22 00:06:07 2022  tmp

Has anyone found a solution for this? I installed the latest chart too without much luck...

[gentoo9ball]

Same problem here

[gentoo9ball]

Did some more research... turns out this was the livenessProbe killing the container because rsync over a sync NFS was taking too long. Reference issue #10

Switching the nfs mount to async reduced time from >20 minutes to 10 seconds.