Game Server Stuck In Pending State, Resource Labeled “Unschedulable”

[GreggoryAddison]

Hello, thanks for taking the time on my post.

I followed this documentation on the Agones website that takes you through a step by step guide on getting a Kubernetes cluster up and running. I set up an auto pilot cluster in the central region. I then used this game server specific documentation  which shows you how to create a custom game server resource. I got back all of the correct responses when walking through the steps. However, when I tried to run my server the console reported back that the resource was unschedulable. I took a look at my quota page and somehow I was using 500/500 gb of my persistent disk. I’m not sure how. There is very little information on how to actually get your own game’s server up and running in GCP so I’ve been patching together documentation. Can anyone link me to a guide detailing the steps on how to get my Unreal Engine Linux Server hosted on a Google Kubernetes cluster using Agones.

STEPS IVE GONE THROUGH 

• Built Unreal Engine Linux Server
• Using Docker, Built a container and Created an Image for the server
• Using GitLab, Pushed that image to a repo
• Using GitLab, created a deploy token for Kubernetes
• In Google Console, created an auto pilot cluster in US-Central1 region
• In Google Console, Verified my credentials 
• In Google Console, added Pull Secret using deploy token from GitLab
• In Google Console, created a custom Game server resource using yaml ( 1GiB memory, 1000m vCpu)
• In Google Console, using kubectl create -f command created game server resource. I get message from console that the resource was created
• In Google Console, using get gameserver command returns my server in the starting status. After 10 min I check the workload and finally saw that it couldn’t be scheduled and the Learn More hyper link for the error took me to a page about quotas.

I’m not a DevOps guy so this stuff is all new to me. Any guidance would be very appreciated! 

[andreicojo5]

Hello @GreggoryAddison ,

From what i understand you have a quota issue, i'm not sure about the limits that you have in place. What i would try is to check inside the pod the storage usage to see if you really are using all the 500GB if so you can try to increase the quota for Persistent Disk Standard (GB). I'm also not sure if it's a quota in kubernetes that you are talking about or a google cloud quota. Also if you could post the exact error message that you get from the pod that would be helpfull.

Best Regards,

Andrei

[garisingh]

Do you recall if the persistent disk quota was for "Persistent Disk SSD"?

Autopilot nodes use PD SSD for their boot disks, so it's possible that you might have hit that limit.