Deploying Vanilla Minecraft Server to Kubernetes

[unruha]

I'm having some trouble deploying the Minecraft server container to Azure Kubernetes Service. I used the Kubernetes manifest file in the examples folder here.

The server seems to start up correctly but then fails when it tries to ping localhost:25565.

Here are the logs when the server is starting...

[init] Running as uid=1000 gid=1000 with /data as 'drwxrwxrwx 6 1000 1000 4096 Jul  9 00:24 /data'
[init] Resolved version given LATEST into 1.19 and major version 1.19
[init] Resolving type given VANILLA
[init] server.properties already created, skipping
[init] Checking for JSON files.
[init] Setting initial memory to 1G and max to 1G
[init] Starting the Minecraft server...
Starting net.minecraft.server.Main
[00:34:15] [ServerMain/INFO]: Building unoptimized datafixer
[00:34:16] [ServerMain/INFO]: Environment: authHost='https://authserver.mojang.com', accountsHost='https://api.mojang.com', sessionHost='https://sessionserver.mojang.com', servicesHost='https://api.minecraftservices.com', name='PROD'
[00:34:18] [ServerMain/INFO]: Loaded 7 recipes
[00:34:18] [ServerMain/INFO]: Loaded 1179 advancements
[00:34:20] [ServerMain/WARN]: Failed to parse level-type default, defaulting to minecraft:normal
[00:34:20] [Server thread/INFO]: Starting minecraft server version 1.19
[00:34:20] [Server thread/INFO]: Loading properties
[00:34:20] [Server thread/INFO]: Default game type: SURVIVAL
[00:34:20] [Server thread/INFO]: Generating keypair
[00:34:20] [Server thread/INFO]: Starting Minecraft server on *:25565
[00:34:20] [Server thread/INFO]: Using epoll channel type
[00:34:20] [Server thread/INFO]: Preparing level "world"
2022-07-09T00:34:24.181Z        INFO    mc-server-runner        Sending 'stop' to Minecraft server...
2022-07-09T00:34:24.182Z        INFO    mc-server-runner        Waiting for completion...

And here is the message when it tries to ping the server and fails...

Events:
  Type     Reason     Age                From               Message
  ----     ------     ----               ----               -------
  Normal   Scheduled  12m                default-scheduler  Successfully assigned default/mc-vanilla-7b67b8ffb-4vxlj to aks-nodepool1-36525092-vmss000001
  Normal   Pulled     12m                kubelet            Successfully pulled image "mchostingacr.azurecr.io/itzg/minecraft-server" in 355.163803ms
  Normal   Killing    11m                kubelet            Container mc-vanilla failed liveness probe, will be restarted
  Warning  Unhealthy  11m (x6 over 11m)  kubelet            Readiness probe failed: WARNING: mcstatus is deprecated; calling mc-monitor instead
server not ready localhost:25565
  Normal   Pulled     11m                kubelet  Successfully pulled image "mchostingacr.azurecr.io/itzg/minecraft-server" in 829.90644ms
  Normal   Started    11m (x2 over 12m)  kubelet  Started container mc-vanilla
  Normal   Created    11m (x2 over 12m)  kubelet  Created container mc-vanilla
  Warning  Unhealthy  11m (x4 over 11m)  kubelet  Liveness probe failed: WARNING: mcstatus is deprecated; calling mc-monitor instead
failed to ping localhost:25565 : could not connect to Minecraft server: dial tcp [::1]:25565: connect: connection refused
  Warning  Unhealthy  10m (x6 over 11m)  kubelet  Readiness probe failed: WARNING: mcstatus is deprecated; calling mc-monitor instead
failed to ping localhost:25565 : could not connect to Minecraft server: dial tcp [::1]:25565: connect: connection refused
  Warning  BackOff  7m2s                kubelet  Back-off restarting failed container
  Normal   Pulling  117s (x8 over 12m)  kubelet  Pulling image "mchostingacr.azurecr.io/itzg/minecraft-server"

The pod/service restarts continuously with the same error.

Does someone with Kubernetes experience know whats going on? Thanks.

[itzg]

It looks like I need to update that example.

What's happening with yours is that the server hasn't fully started up in time, so it then fails the pod. Here's what I use lately in my deployment :

          readinessProbe:
            exec:
              command: ["/usr/local/bin/mc-monitor", "status", "--host", "localhost"]
            initialDelaySeconds: 20
            periodSeconds: 10
            failureThreshold: 12

[unruha]

This worked for me, thanks a lot. However I had to increase initialDelaySeconds to 60. Probably because I'm running on slower nodes.