Artifactory fails to start due to "unbound immediate PersistentVolumeClaims". PVC however is bound. Longhorn storage.

[serrnovik]

Context:

• kubernetes: v1.22.4
• chart: jfrog/artifactory-oss

With:

• No customizations, just: helm upgrade --install artifactory --namespace artifactory jfrog/artifactory-oss;
• On the cluster backed by 'longhorn' storage;
• No other StorrageClasses configured in the cluster.

Problem

Artifactory' pod ends up in error state:

0/5 nodes are available: 5 pod has unbound immediate PersistentVolumeClaims.
Back-off restarting failed container

5 is the number of nodes in the cluster

pod is never created, so no more logs.

However:

At the same time there are 2 PVC bound:

Postgresql pod is up and healthy, volume bound and used.

Expected:

Artifactory pod uses bound volume and goes up like Postgresql one.

---

seems that other people had problems having longhorn storage backend: #1610

[uakharkivponomarov]

I have encountered a similar problem. It arises when I install the Artifactory chart (located at /stable/artifactory) from scratch with splitServicesToContainers: true. The Artifactory deployment succeeds only after 3 attempts. Please find the attachments with the StatefulSet pod events for each attempt and persistent volume claim is left behind after the first removing the Helm release.

Conversely, when splitServicesToContainers is set to false, the deployment from scratch completes without any errors.
run-attemp-1_artifactory-0.log
run-attemp-2_artifactory-1.log