Build: Creating a deployment: Cannot deploy docker container from control plane node

blaket

Hello, I'm on the Build section, trying to deploy an instance of my newly created image to my control plane.
when I run the kubectl create deployment command to create the image, it says
deployment.apps "simpleappdeploy" created
but then when I check my pods with kubectl get pods I get back
simpleappdeploy-7b7686dcdb-6676d 0/1 CrashLoopBackOff 1 (6s ago) 7s
I'm not sure what is going wrong.
I am able to deploy the container successfully locally on my local laptop using the exact same kubectl command, but for some reason the control plane node made by the specs in this class fails to get the simpleapp to run.
Does anyone have any suggestions for what I could check that could be causing this simpleapp to fail?
It is the app that just creates a log entry to a file every 5 seconds.

blaket

If it helps, when I run the kubectl describe command for the broken pod, the output is

Name: simpleappdeploy-7b7686dcdb-grbqg
Namespace: default
Priority: 0
Service Account: default
Node: ip-172-31-30-3/172.31.30.3
Start Time: Thu, 11 May 2023 15:40:32 +0000
Labels: app=simpleappdeploy
pod-template-hash=7b7686dcdb
Annotations: cni.projectcalico.org/containerID: c4266db26f5f7105644c12f6735daf610c12373983fc5271fa5a2c97f9b0d018
cni.projectcalico.org/podIP: 192.168.37.211/32
cni.projectcalico.org/podIPs: 192.168.37.211/32
Status: Running
IP: 192.168.37.211
IPs:
IP: 192.168.37.211
Controlled By: ReplicaSet/simpleappdeploy-7b7686dcdb
Containers:
simpleapp:
Container ID: containerd://bf3f627cfe095f804d63c9965bcf8cc5d0bb9fc290c38983600bf3baf72b680e
Image: blakethompson/simpleapp:1.0
Image ID: docker.io/blakethompson/[email protected]:220983840b4fd5bf28ef80c47f0d445bab2bf2c3150aca52cb296dcbade04ba6
Port:
Host Port:
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Error
Exit Code: 1
Started: Thu, 11 May 2023 15:40:46 +0000
Finished: Thu, 11 May 2023 15:40:46 +0000
Ready: False
Restart Count: 2
Environment:
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-4zzql (ro)
Conditions:
Type Status
Initialized True
Ready False
ContainersReady False
PodScheduled True
Volumes:
kube-api-access-4zzql:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional:
DownwardAPI: true
QoS Class: BestEffort
Node-Selectors:
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 34s default-scheduler Successfully assigned default/simpleappdeploy-7b7686dcdb-grbqg to ip-172-31-30-3
Normal Pulled 20s (x3 over 34s) kubelet Container image "blakethompson/simpleapp:1.0" already present on machine
Normal Created 20s (x3 over 34s) kubelet Created container simpleapp
Normal Started 20s (x3 over 33s) kubelet Started container simpleapp
Warning BackOff 5s (x4 over 32s) kubelet Back-off restarting failed container simpleapp in pod simpleappdeploy-7b7686dcdb-grbqg_default(4461be56-558a-46be-b146-6aac9a722c72)

And when I run kubectl logs {pod} the output is
exec /usr/local/bin/python: exec format error

chrispokorni

Hi @blaket,

What infrastructure are you using for lab VMs, and did you follow all resources available in the course material (Infrastructure provisioning videos for AWS and GCP, and lab guide instructions) to properly configure any related services, such as VPC, SG, firewall, VM instance size, recommended guest OS, recommended Kubernetes version, etc?

If the deployment is successful on one Kubernetes environment (seems local from your description), but unsuccessful on another (cloud maybe?), it is most likely a configuration difference between the two clusters that allows/prevents the container from running respectively.

Regards,
-Chris

blaket

Hi @chrispokorni, I think that I found the issue!
I wasn't originally able to install docker using the sudo apt-install command, so I just made the docker image locally on my Mac.
The problem was that when I pushed that image up to my dockerhub repository, the image then had a different platform from what the Ubuntu server uses.
I went through the instructions here to install docker onto the ubuntu machine, recreated the image there, pushed the image back to the repository, and then found that I was able to successfully deploy the image to my worker node

chrispokorni

Hi @blaket,

Ah yes, the new ARM issues that seem to be frequently reported...

Regards,
-Chris