Container Compute Service:Configure scheduling policies to ignore the resource requests of specific sidecar containers

Feature description

ACS clusters do not support DaemonSets due to the limits of virtual nodes. To achieve a DaemonSet-like effect, you can add sidecar containers to pods in ACS clusters. In addition, ACS allows you to ignore the resource requests of a sidecar container by adding an environment variable to the configurations of the container. This way, when the ACS cluster adjusts and schedules resources, the system ignores the resource requests of the sidecar container. When the system runs the pod, the system allocates resources to the pod based on the resources.requests and resources.limits parameter of other containers in the pod. This ensures resource sharing and isolation between the sidecar container and other containers in the pod.

Configuration description

To ignore the resource requests of a container, add the __IGNORE_RESOURCE__ environment variable to the container configuration and set the value to true.

Sample configuration

1. Create a file named test-ignore.yaml and copy the following content to the file:

   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: test-ignore
     labels:
       app: test
   spec:
     replicas: 1 
     selector:
       matchLabels:
         app: test
     template:
       metadata:
         labels:
           app: test
       spec:
         containers:
         - name: app
           image: registry-cn-hangzhou.ack.aliyuncs.com/ack-demo/busybox:1.28
           command: ["/bin/sh", "-c", "sleep 999"]
           resources:
             requests:
               cpu: 1
               memory: 2Gi
         - name: sidecar
           image: registry-cn-hangzhou.ack.aliyuncs.com/ack-demo/busybox:1.28
           command: ["/bin/sh", "-c", "sleep 999"]
           resources:
             requests:
               cpu: 0.25
               memory: 0.5Gi

   This YAML file is used to create a Deployment that contains one replicated pod. The pod contains two containers, each of which is configured with resource requests.

2. Run the following command to create the Deployment:

   kubectl apply -f test-ignore.yaml

3. Run the following command to query information about the pod and view the resource configurations of the pod after adjustment:

   kubectl get pod <pod-name> -o yaml | grep alibabacloud.com/pod-use-spec

   Expected output:

   alibabacloud.com/pod-use-spec: 2-4Gi

   The output shows that the pod after adjustment occupies 2 vCPUs and 4 GiB of memory when the resource requests of the sidecar container are not ignored.

4. Modify the sidecar container configurations in the test-ignore.yaml file based on the following code block. The following code block adds an environment variable to ignore the resource requests of the sidecar container.

   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: test-ignore
   .....
         - name: sidecar
           image: registry-cn-hangzhou.ack.aliyuncs.com/ack-demo/busybox:1.28
           command: ["/bin/sh", "-c", "sleep 999"]
           env:
           - name: __IGNORE_RESOURCE__
             value: "true"
           resources:
             requests:
               cpu: 0.25
               memory: 0.5Gi

5. Run the following command to update the Deployment:

   kubectl apply -f test-ignore.yaml

6. Run the following command again to query information about the pod and view the resource configurations of the pod after adjustment:

   kubectl get pod <pod-name> -o yaml | grep alibabacloud.com/pod-use-spec

   The output shows that the pod after adjustment occupies 1 vCPU and 2 GiB of memory when the resource requests of the sidecar container are ignored.

7. Check the CPU weights of the running containers.

   1. Run the following command to log on to the app container and query the CPU weight of the app container:

      kubectl exec -ti <pod-name> -c app -- cat /sys/fs/cgroup/cpu/cpu.shares

      Expected output:

      1024

   2. Run the following command to log on to the sidecar container and query the CPU weight of the sidecar container:

      kubectl exec -ti <pod-name> -c sidecar -- cat /sys/fs/cgroup/cpu/cpu.shares

      Expected output:

      256

      The output shows that the ratio of the CPU shares used by the app container to the CPU shares used by the sidecar container is 4:1.

The following table describes the resource adjustment and resource isolation results for different pod resource configurations.