fix bugs that postgres can't start

[sunlianqiang]

after k8s node restart, the postgres of harbor can't restart. It's file mode issue when checking logs

kubectl logs -f harbor-sh-harbor-database-0 -n harbor-sh
FATAL:  data directory "/var/lib/postgresql/data" has group or world access
DETAIL:  Permissions should be u=rwx (0700).

I find a method that change the config of initContainers to fix the bug. I Also found in https://github.com/helm/charts Postgres are configured with master-slave, hoping to use asap, I also can help if is needed.

FYI:
ClusterHQ/dvol#45
https://github.com/helm/charts/tree/master/stable/postgresql

[sunlianqiang]

@ywk253100 @heww 大佬，帮忙review下？

[ywk253100]

Please refer to goharbor/harbor#8224 and let's try to fix it in the entrypoint script so that the solution can apply to both OSS Harbor and Harbor chart

[yydzhou]

Is there any blocking issue to let this MR in? Or the problem resolved in other ways?

[jbrandli]

I believe the issue was solved in another PR and issue already  On Wednesday, July 29, 2020, 04:21:09 PM MDT, yydzhou <[email protected]> wrote: Is there any blocking issue to let this MR in? Or the problem resolved in other ways? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[sunlianqiang]

I believe the issue was solved in another PR and issue already  On Wednesday, July 29, 2020, 04:21:09 PM MDT, yydzhou [email protected] wrote: Is there any blocking issue to let this MR in? Or the problem resolved in other ways? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

so it's no need for this pr any more, close it now

[yydzhou]

I believe the issue was solved in another PR and issue already  On Wednesday, July 29, 2020, 04:21:09 PM MDT, yydzhou [email protected] wrote: Is there any blocking issue to let this MR in? Or the problem resolved in other ways? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

so it's no need for this pr any more, close it now

I believe the issue was solved in another PR and issue already  On Wednesday, July 29, 2020, 04:21:09 PM MDT, yydzhou [email protected] wrote: Is there any blocking issue to let this MR in? Or the problem resolved in other ways? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

It's not resolved at all. I have tried the most recent version and the issue still exists in both v1.3.4 and v1.4.2. I don't know where the conclusion of "resolved" comes from.

[couillonnade]

It's not resolved. Chmod still did the trick for me in the latest version (1.4.2)

[romnn]

I can confirm the same behaviour as @yydzhou with current versions in minikube when using fresh PVC's.

[gotiniens]

I can confirm the same behaviour with the latest (1.5.0) helm chart.

I can reproduce the error as follows:

*install harbor with helm chart, using a rook-ceph RBD volume as storage backend for the database
*downscale the database deployment to 0
*wait for the deployment to be terminated
*upscale to 1

[ywk253100]

Please refer to #444 (comment)