Could not determine my container ID: Container not found

Portworx Install
1

Hi all,

I’m trying to install portworx on a rke2 cluster provisionned on vsphere using rancher 2. I’ve tried using the helm chart and px-central but each time it fails with the same error. The portworx pods fail to start with these errors :

time=“2022-02-04T15:57:41Z” level=error msg=“Could not instantiate container client” error=“Could not determine my container ID: Container not found”
Fri, Feb 4 2022 4:57:41 pm time=“2022-02-04T15:57:41Z” level=error msg=“Could not talk to Docker/Containerd/CRI: Could not determine my container ID: Container not found - please ensure ‘/var/run/docker.sock’, ‘/run/containerd/containerd.sock’ or ‘/var/run/crio/crio.sock’ are mounted.”

I took a look at the daemonset spec, the k3s socket (/run/k3s/containerd/containerd.sock) is mounted on /run/containerd/containerd.sock, so it should talk with containerd without any problem. I tried to add --debug to the command argument but not very helpfull. What could cause this error ?
FYI, I tried an installation on rke1 and had the same error…

Best regards

2

I’m getting a similar error when installing in okd 4.9 with the px-cluster pods. Volumes are all there and appear to be mounted.

level=error msg=“Could not talk to Docker/Containerd/CRI: Could not determine my container ID: Container not found - please ensure ‘/var/run/docker.sock’, ‘/run/containerd/containerd.sock’ or ‘/var/run/crio/crio.sock’ are mounted.”

3

I would also like to report the same problem with OKD 4.9. Is there any solution for this problem yet?

4

Getting the same issue on Rocky Linux 8.5, Docker version 20.10.12, build e91ed57.
Cannot find the socket inside container, although it exists in the /var/run/ directory.

5

Still having this problem on an OKD 4.9 cluster with Portworx 2.10.

i have also verified that the socket is present in the container.

6

@ bsauvajon use the operator spec instead of daemonset. can you please post your install spec yaml file?

7

Hi,

I’ve made some tests using operator with the latest version on k3s 1.24.3 and the problem is still here.
My StorageCluster definition is very basic :

kind: StorageCluster
apiVersion: core.libopenstorage.org/v1
metadata:
  name: px-cluster-c113084f-89ff-4aee-b178-797d9d729846
  namespace: kube-system
  annotations:
    portworx.io/install-source: "https://install.portworx.com/?operator=true&mc=false&oem=esse&user=ab173b78-fbad-4d05-92d1-1eda3ca24126&b=true&s=%2Fdev%2Fdata-vg%2Fpx-data%2C%2Fdev%2Fdata-vg%2Fpx-jrnal&kd=%2Fdev%2Fdata-vg%2Fpx-kvdb&c=px-cluster-c113084f-89ff-4aee-b178-797d9d729846&stork=true&csi=true&mon=true&tel=false&st=k8s&promop=true"
    portworx.io/misc-args: "--oem esse"
spec:
  image: portworx/oci-monitor:2.10.3
  imagePullPolicy: Always
  kvdb:
    internal: true
  storage:
    devices:
    - /dev/data-vg/px-data
    journalDevice: /dev/data-vg/px-jrnal
    kvdbDevice: /dev/data-vg/px-kvdb
  secretsProvider: k8s
  stork:
    enabled: true
    args:
      webhook-controller: "true"
  autopilot:
    enabled: true
  monitoring:
    prometheus:
      enabled: true
      exportMetrics: true
  featureGates:
    CSI: "true"

So I don’t think the problem comes from this spec