No logs received from other component once it has been redeployed

[jellevanhees]

Describe the bug Fluentd runs fine and logs everything at startup. but when other containers get new deploys the logs stop showing up in elasticsearch. once we restart fluentd they show up again.

Version of Helm and Kubernetes:

Helm Version:

$ helm version
version.BuildInfo{Version:"v3.4.2", GitCommit:"23dd3af5e19a02d4f4baa5b2f242645a1a3af629", GitTreeState:"clean", GoVersion:"go1.14.13"}

Kubernetes Version:

AWS EKS cluster nodes are AWS managed nodes groups al default.

$ kubectl version
Client Version: version.Info{Major:"1", Minor:"20", GitVersion:"v1.20.0", GitCommit:"af46c47ce925f4c4ad5cc8d1fca46c7b77d13b38", GitTreeState:"clean", BuildDate:"2020-12-08T17:59:43Z", GoVersion:"go1.15.5", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"19+", GitVersion:"v1.19.6-eks-49a6c0", GitCommit:"49a6c0bf091506e7bafcdb1b142351b69363355a", GitTreeState:"clean", BuildDate:"2020-12-23T22:10:21Z", GoVersion:"go1.15.5", Compiler:"gc", Platform:"linux/amd64"} 

Which version of the chart: 11.9.0

What happened: once another deployment cycles after a new version of it is deployed logs stop showing up

What you expected to happen: I can redeploy other tools as much as I like without losing logs

How to reproduce it (as minimally and precisely as possible): this is in a pretty minimal setup. we have a basic chart that depends on version 11.9.0 nothing else is installed by this chart. other components in the cluster are also installed via helm charts

values.yaml (only put values which differ from the defaults)

fluentd-elasticsearch:
  image:
    pullPolicy: IfNotPresent

  # Specify where fluentd can find logs
  hostLogDir:
    varLog: /var/log
    dockerContainers: /var/lib/docker/containers
    libSystemdDir: /usr/lib64

  elasticsearch:
    hosts:
      - elasticnode.internal:80
    scheme: 'http'
    ssl_version: TLSv1_2
    auth:
      enabled: true
      user: "usernam"
      password: "password"
    logstash:
      enabled: true
      prefix: "eksstaging"
    buffer:
      enabled: true
      # ref: https://docs.fluentd.org/configuration/buffer-section#chunk-keys
      chunkKeys: ""
      type: "file"
      path: "/var/log/fluentd-buffers/kubernetes.system.buffer"
      flushMode: "interval"
      retryType: "exponential_backoff"
      flushThreadCount: 2
      flushInterval: "5s"
      retryForever: true
      retryMaxInterval: 30
      chunkLimitSize: "256M"
      queueLimitLength: 20
      overflowAction: "block"
  # If you want to change args of fluentd process
  # by example you can add -vv to launch with trace log
  fluentdArgs: "--no-supervisor -q"

install command

helm upgrade --install ops-fluentd-elasticsearch . --namespace ops-fluentd-elasticsearch -f acc-values.yaml

chart.yml

apiVersion: v1
appVersion: "1.0"
description: A Helm chart for Kubernetes
name: fluentd-elasticsearch
version: 11.9.0
dependencies:
  - name: fluentd-elasticsearch
    version: 11.9.0
    repository: "@kokuwa"
    condition:
    tags:

Anything else we need to know:

[nickjmv]

We have the exact same problem in our team.

[zifeo]

@monotek How could we help debugging this?