API server not reachable when using redis cache

[jasondamour]

Hello,

I have a running instance of chartmuseum that is working (v0.13.0 running on Kubernetes, GCS backend, local cache). When adding Redis as a cache, the container stops serving on 8080 even though the logs indicate the process is healthy.

Here is the change which breaks the container:

          env:
+         - name: "CACHE"
+           value: "redis"
+         - name: "CACHE_REDIS_ADDR"
+           value: "10.76.0.3:6379"
+         - name: "DISABLE_STATEFILES"
+           value: "true"
+         - name: "CACHE_REDIS_PASSWORD"
+           valueFrom:
+             secretKeyRef:
+               name: chartmuseum-chartmuseum
+               key: "CACHE_REDIS_PASSWORD"

and in the logs, I do see a successful connection (repeating as the container gets restarted again and again):

Entry found in cache store {"repo": ""}
Fetching chart list from storage {"repo": ""}
Entry found in cache store {"repo": ""}
Fetching chart list from storage {"repo": ""}
Entry found in cache store {"repo": ""}
Fetching chart list from storage {"repo": ""}
Entry found in cache store {"repo": ""}
Fetching chart list from storage {"repo": ""}```

However, just this change causes all health/liveness probes to fail:

Liveness probe failed: HTTP probe failed with statuscode: 500

and I'm unable to tunnel to port 8080 of the container (connection refused) even after disabling the probes to let the container run uninterrupted

[nerdeveloper]

Thank you for this. We are going to look into this. @phirmware you can take a look at this as your first issue

[phirmware]

Seen, will look into this.

[jasondamour]

Works now