changing type to SPIGOT causes server to stop running.

[damienj210]

I am trying to run the standard minecraft server helm chart, I choose to customize helm options before install. These are the items I change:

eula: 'TRUE'
motd: testing Minecraft on Kubernetes!
onlineMode: false
type: SPIGOT
persistence:
  annotations: {}
  dataDir:
    Size: 10Gi
    enabled: true
    existingClaim: pvc-testminecraft

Watching the logs during the pod creation, everything seems to be ok and it completes with "Done" but minutes later it shuts down the server. I have attached my logs and yaml files.

testminecraft-minecraft-857d4fd99b-mtbbm_testminecraft-minecraft.log testminecraft.yaml.txt

Please let me know what I'm doing wrong. This only happens when I change the type to SPIGOT. If I make all of the other changes but leave type as VANILLA it works fine.

Thanks.

[itzg]

Something signaled the pod/container to stop, so it did a graceful stop:

2024-01-24T02:37:39.671Z    INFO    mc-server-runner    Sending 'stop' to Minecraft server...
2024-01-23T20:37:39.671988020-06:00 2024-01-24T02:37:39.671Z    INFO    mc-server-runner    Waiting for completion...
[02:37:39] [Server thread/INFO]: Stopping the server
[02:37:39] [Server thread/INFO]: Stopping server
[02:37:39] [Server thread/INFO]: Saving players
[02:37:39] [Server thread/INFO]: Saving worlds
[02:37:45] [Server thread/INFO]: Saving chunks for level 'ServerLevel[world]'/minecraft:overworld

Perhaps use kubectl get events to see what kubernetes operation initiated it.

[panromir]

Hey there :)

I have run into the same issue and I assume it's the Probes.

I replaced them in the Helm chart values.yaml and that fixes the issue I can now play MC fine.

livenessProbe:
  enabled: false
  command:
    - echo
  initialDelaySeconds: 30
  periodSeconds: 5
  failureThreshold: 20
  successThreshold: 1
  timeoutSeconds: 1
readinessProbe:
  enabled: false
  command:
    - echo
  initialDelaySeconds: 30
  periodSeconds: 5
  failureThreshold: 20
  successThreshold: 1
  timeoutSeconds: 1
startupProbe:
  command:
    - echo
  enabled: false
  failureThreshold: 30
  periodSeconds: 10

Of course, that's not ideal!

Now the original issue is the following:

The mc-status script errors out:

minecraft-minecraft-6cf79f486b-ftwtp:/data$ mc-health
failed to ping localhost:25565 : json: cannot unmarshal string into Go struct field ServerInfo.description of type mcpinger.RegularChatComponent

What mc-monitor does I don't know, since it's a go binary. This happens with Spigot 1.20.4

[itzg]

@panromir that latest behavior is being tracked in https://github.com/itzg/docker-minecraft-server/issues/2545 and will be fixed soon

[panromir]

@panromir that latest behavior is being tracked in itzg/docker-minecraft-server#2545 and will be fixed soon

This has fixed the Probes and the server is now running properly in k3s with Spigot 1.20.4 with the original values.yaml configuration. Thanks!

[itzg]

Resolved by https://github.com/itzg/docker-minecraft-server/pull/2628