set maxprocs error: invalid format for CGroupSubsys

[dongjiang1989]

What version of Loggie?

v1.4.0

Expected Behavior

2023-03-07 19:10:11 INF cmd/loggie/main.go:69 > version: v1.4.0-33d018f
2023-03-07 19:10:11 FTL cmd/loggie/main.go:76 > set maxprocs error: invalid format for CGroupSubsys: "12:pids:/system.slice/containerd.service/kubepods-pod6392e259_4635_4667_a568_3cc95ecce1eb.slice:cri-containerd:93d73b291fa5e0ec74f1c7673ab4aebb38210e8a0012227d922cba55dd614eb5"

Actual Behavior

loggie DaemonSets deploy to k8s

Steps to Reproduce the Problem

[ethfoo]

I don't know if it is a problem with the "go.uber.org/automaxprocs/maxprocs", please provide your environment information, such as the operating system version, the container runtime, the k8s version, etc.

[imlense]

provide

Kubernetes ： v1.21.5 OS: centos 8.2 Kernal: 4.19.0-240.23.21.el8_2 containerd: v1.5.7

loggie daemonSet yaml：

---
# Source: loggie/templates/loggie-rbac.yaml
apiVersion: v1
kind: ServiceAccount
metadata:
  name: loggie
  namespace: monitoring
---
# Source: loggie/templates/loggie-configmap.yaml
apiVersion: v1
data:
  loggie.yml: |-
    loggie:
      defaults:
        sink:
          type: loki
          url: http://loki.monitoring:3100/loki/api/v1/push
      discovery:
        enabled: true
        kubernetes:
          containerRuntime: containerd
          dynamicContainerLog: true
          parseStdout: false
          rootFsCollectionEnabled: true
          typeNodeFields:
            clusterlogconfig: ${_k8s.clusterlogconfig}
            nodename: ${_k8s.node.name}
            os: ${_k8s.node.nodeInfo.osImage}
          typePodFields:
            containername: ${_k8s.pod.container.name}
            logconfig: ${_k8s.logconfig}
            namespace: ${_k8s.pod.namespace}
            nodename: ${_k8s.node.name}
            podname: ${_k8s.pod.name}
      http:
        enabled: true
        port: 9196
      monitor:
        listeners:
          filesource:
            period: 10s
          filewatcher:
            period: 5m
          pipeline:
            period: 10s
          queue:
            period: 10s
          reload:
            period: 10s
          sink:
            period: 10s
        logger:
          enabled: true
          period: 30s
      reload:
        enabled: true
        period: 10s
kind: ConfigMap
metadata:
  name: loggie-config-loggie
  namespace: monitoring
---
# Source: loggie/templates/loggie-rbac.yaml
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: loggie-role-loggie
rules:
  - apiGroups:
    - ""
    resources:
      - pods
      - pods/log
      - nodes
    verbs:
      - get
      - watch
      - list
  - apiGroups:
    - ""
    resources:
      - events
    verbs:
      - get
      - watch
      - list
      - update
      - create
      - patch
  - apiGroups:
    - extensions
    - apps
    resources:
      - deployments
    verbs:
      - get
      - list
      - watch
  - apiGroups:
    - loggie.io
    resources:
      - logconfigs
      - logconfigs/status
      - clusterlogconfigs
      - clusterlogconfigs/status
      - sinks
      - interceptors
    verbs:
      - get
      - list
      - watch
      - update
      - patch
  - apiGroups:
    - coordination.k8s.io
    resources:
      - leases
    verbs:
      - create
      - get
      - list
      - update
  - apiGroups:
    - metrics.k8s.io
    resources:
      - pods
    verbs:
      - get
      - list
---
# Source: loggie/templates/loggie-rbac.yaml
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: loggie-rolebinding-loggie
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: loggie-role-loggie
subjects:
  - kind: ServiceAccount
    name: loggie
    namespace: monitoring
---
# Source: loggie/templates/loggie-service.yaml
apiVersion: v1
kind: Service
metadata:
  labels:
    app: loggie
    instance: loggie
  name: loggie
  namespace: monitoring
spec:
  ports:

    - name: monitor
      port: 9196
      targetPort: 9196
  selector:
    app: loggie
    instance: loggie
  type: ClusterIP
---
# Source: loggie/templates/loggie-agent-daemonset.yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  labels:
    app: loggie
    instance: loggie
  name: loggie
  namespace: monitoring
spec:
  selector:
    matchLabels:
      app: loggie
      instance: loggie
  template:
    metadata:
      labels:
        app: loggie
        instance: loggie
    spec:
      containers:
        - args:
            - -meta.nodeName=$(HOST_NAME)
            - -config.system=/opt/loggie/loggie.yml
            - -config.pipeline=/opt/loggie/pipeline/*.yml
            - -log.jsonFormat=true
          env:
            - name: HOST_NAME
              valueFrom:
                fieldRef:
                  apiVersion: v1
                  fieldPath: spec.nodeName
            - name: TZ
              value: Asia/Shanghai
          image: loggieio/loggie:v1.4.0
          name: loggie
          resources:
            limits:
              cpu: 2
              memory: 2Gi
            requests:
              cpu: 100m
              memory: 100Mi
          volumeMounts:

            - mountPath: /var/log/pods
              name: podlogs
            - mountPath: /var/lib/docker/containers
              name: dockercontainers
            - mountPath: /var/lib/kubelet/pods
              name: kubelet
            - mountPath: /opt/loggie/loggie.yml
              name: loggie-config
              subPath: loggie.yml
            - mountPath: /opt/loggie/pipeline
              name: pipeline
            - mountPath: /data/
              name: registry
            - mountPath: /run/containerd/containerd.sock
              name: containerdsocket

      serviceAccountName: loggie
      nodeSelector:
        {}
      affinity:
        {}
      tolerations:
        - effect: NoSchedule
          key: node-role.kubernetes.io/master
          operator: Exists
        - effect: NoSchedule
          key: node-role.kubernetes.io/control-plane
          operator: Exists
      volumes:
        - hostPath:
            path: /var/log/pods
            type: DirectoryOrCreate
          name: podlogs
        - hostPath:
            path: /var/lib/docker/containers
            type: DirectoryOrCreate
          name: dockercontainers
        - hostPath:
            path: /var/lib/kubelet/pods
            type: DirectoryOrCreate
          name: kubelet
        - configMap:
            defaultMode: 384
            name: loggie-config-loggie
          name: loggie-config
        - hostPath:
            path: /data/loggie-loggie
            type: DirectoryOrCreate
          name: registry
        - emptyDir: {}
          name: pipeline
        - hostPath:
            path: /run/containerd/containerd.sock
            type: ""
          name: containerdsocket
      hostPID: true

  updateStrategy:
    type: RollingUpdate
---
# Source: loggie/templates/loggie-servicemonitor.yaml
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: loggie-config-loggie
  namespace: monitoring
  labels:
    app: loggie
    instance: loggie
    release: metrics
spec:
  jobLabel: jobLabel
  endpoints:
  - port: monitor
    interval: 30s
    honorLabels: true
  selector:
    matchLabels:
      app: loggie
      instance: loggie

[ethfoo]

I feel that the cgoup v2 version will cause this problem, because the lower version of automaxprocs does not support it. refer: https://github.com/uber-go/automaxprocs/releases/tag/v1.5.0

@dongjiang1989 Could you Please change the content of the PR to upgrade the version of automaxprocs and try again?

[dongjiang1989]

update v1.5.1 fix bug

fixs #488