kubernetes 为什么气流花吊舱的活/就绪探测失败？

svujldwt 于 2023-03-07 发布在 Kubernetes

关注(0)|答案(1)|浏览(150)

我正在尝试通过Helm图表apache-airflow/airflow:1.5.0将Airflow部署到Kubernetes群集。airflow-flower-* pod中出现错误：

Readiness probe errored: rpc error: code = Unknown desc = failed to exec in container: failed to start exec "23ece1d0a3bf7e9dc06f8e70cfb0e1d69687e2cbea952f18da7b4e80a2211995": OCI runtime exec failed: exec failed: unable to start container process: error executing setns process: exit status 1: unknown

Readiness probe failed: % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0curl: (7) Failed to connect to localhost port 5555: Connection refused

Liveness probe failed: % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0curl: (7) Failed to connect to localhost port 5555: Connection refused

探测器是curl localhost:5555。我的override-values.yaml看起来像：

extraSecrets:
  airflow-postgresql-metadata:
    data: "$AIRFLOW_POSTGRESQL_METADATA"
  airflow-postgresql-result-backend:
    data: "$AIRFLOW_POSTGRESQL_RESULT_BACKEND"

data:
  metadataSecretName: airflow-postgresql-metadata
  resultBackendSecretName: airflow-postgresql-result-backend

workers:
  persistence:
    enabled: false
  resources:
    limits:
     cpu: 100m
     memory: 128Mi
    requests:
     cpu: 100m
     memory: 128Mi
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

scheduler:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

webserver:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

flower:
  resources:
    limits:
      cpu: 100m
      memory: 128Mi     
    requests:
      cpu: 100m
      memory: 128Mi
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

statsd:
  resources:
    limits:
      cpu: 100m
      memory: 128Mi     
    requests:
      cpu: 100m
      memory: 128Mi
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

cleanup:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

createUserJob:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

migrateDatabaseJob:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

triggerer:
  enabled: false
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

rbac:
  create: false

redis:
  persistence:
    storageClassName: "ceph-rbd-ssd"
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa

postgresql:
  enabled: false

也就是说，我几乎什么都没做，我只是设置了资源限制并禁用了服务帐户的创建。
此外，我还对部署AirFlow的values.yaml文件的最小版本感兴趣

kubernetes

来源：https://stackoverflow.com/questions/75601082/why-liveness-readiness-probe-of-airflow-flower-pod-failed

1条答案

按热度按时间

bjp0bcyl1#

这不是问题的答案，而是解决问题
我关闭了花，但我不断从另一个豆荚得到同样的错误。根据豆荚的日志我发现根本没有发生任何事情。我以为是缺乏资源-事实证明
这个values.yaml为我工作：

extraSecrets: 
  airflow-postgresql-metadata:
    data: "$AIRFLOW_POSTGRESQL_METADATA"
  airflow-postgresql-result-backend:
    data: "$AIRFLOW_POSTGRESQL_RESULT_BACKEND"

# Airflow database & redis config
data:
  metadataSecretName: airflow-postgresql-metadata
  resultBackendSecretName: airflow-postgresql-result-backend

# Airflow Worker Config
workers:
  replicas: 1
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  keda:
    enabled: false
  persistence:
    enabled: false
  kerberosSidecar:
    enabled: false
  resources:
   limits:
    cpu: 1000m
    memory: 2048Mi
   requests:
    cpu: 1000m
    memory: 2048Mi
  logGroomerSidecar:
    enabled: false

# Airflow scheduler settings
scheduler:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  resources:
   limits:
    cpu: 1000m
    memory: 2048Mi
   requests:
    cpu: 1000m
    memory: 2048Mi
  logGroomerSidecar:
    enabled: false

# Airflow webserver settings
webserver:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  podDisruptionBudget:
    enabled: false
  resources:
    limits:
      cpu: 1000m
      memory: 2048Mi
    requests:
      cpu: 1000m
      memory: 2048Mi

# Airflow Triggerer Config
triggerer:
  enabled: true
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  resources:
   limits:
    cpu: 1000m
    memory: 2048Mi
   requests:
    cpu: 1000m
    memory: 2048Mi

# StatsD settings
statsd:
  enabled: true
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  resources:
    limits:
      cpu: 1000m
      memory: 2048Mi
    requests:
      cpu: 1000m
      memory: 2048Mi

# Configuration for the redis provisioned by the chart
redis:
  enabled: true
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  persistence:
    enabled: false
  resources:
   limits:
    cpu: 1000m
    memory: 2048Mi
   requests:
    cpu: 1000m
    memory: 2048Mi

# Airflow create user job settings
createUserJob:
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  resources:
    limits:
      cpu: 1000m
      memory: 2048Mi
    requests:
      cpu: 1000m
      memory: 2048Mi

# Airflow database migration job settings
migrateDatabaseJob:
  enabled: true
  serviceAccount:
    create: false
    name: srs-khajiit-airflow-sa
  resources:
    limits:
      cpu: 1000m
      memory: 2048Mi
    requests:
      cpu: 1000m
      memory: 2048Mi

cleanup:
  enabled: false

postgresql:
  enabled: false

pgbouncer:
  enabled: false

dags:
  persistence:
    enabled: false
  gitSync:
    enabled: false

dagProcessor:
  enabled: false

flower:
  enabled: false

logs:
  persistence:
    enabled: false

elasticsearch:
  enabled: false

kerberos:
  enabled: false

ingress:
  enabled: false
  web:
    enabled: false
  flower:
    enabled: false
    
networkPolicies:
  enabled: false

rbac:
  create: false
  createSCCRoleBinding: false

赞(0）回复(0）举报 2023-03-07

我来回答

kubernetes 为什么气流花吊舱的活/就绪探测失败？

1条答案

相关问题

热门标签

最新问答