Fluentd Pod が再起動を繰り返し、CrashLoopBackOff で停止する

Issue

• openshift-logging namespace 内のすべての logging-fluentd-xxxxx Pod がクラッシュし続けます。

logging-fluentd-jj77j                         0/1       CrashLoopBackOff    1131       4d        10.0.0.1     node02.example.com    <none>
logging-fluentd-lwp6z                         0/1       CrashLoopBackOff    549        4d      10.0.0.2    node01.example.com    <none>
logging-fluentd-s9nsp                         0/1       CrashLoopBackOff    834        4d        10.0.0.3    node03.example.com    <none>

• ノード上の Docker サービスログには、以下のエラーメッセージが表示されます。

Aug 22 02:03:17 node1.example.com dockerd-current[17000]: time="2020-08-22T02:03:17.722980022-04:00" level=error msg="Handler for POST /v1.26/containers/create?name=k8s_fluentd-elasticsearch_logging-fluentd-mksxs_openshift-logging_bc70e819-ec22-11e9-b59b-566f2fcb01c7_6332 returned error: Conflict. The container name \"/k8s_fluentd-elasticsearch_logging-fluentd-mksxs_openshift-logging_bc70e819-ec22-11e9-b59b-566f2fcb01c7_6332\" is already in use by container 05a8e05ccbac065ff191ae840f07a21918b26afe30047008eb7dea4d8feef06c. You have to remove (or rename) that container to be able to reuse that name."

• Fluentd Pod ログには以下のエラーメッセージが表示されます。

2020-08-23 11:14:43 -0400 [error]: unexpected error error="getaddrinfo: Name or service not known"
  2020-08-23 11:14:43 -0400 [error]: /usr/share/gems/gems/fluentd-0.12.43/lib/fluent/plugin/out_forward.rb:508:in `getaddrinfo'
  2020-08-23 11:14:43 -0400 [error]: /usr/share/gems/gems/fluentd-0.12.43/lib/fluent/plugin/out_forward.rb:508:in `resolve_dns!'
  2020-08-23 11:14:43 -0400 [error]: /usr/share/gems/gems/fluentd-0.12.43/lib/fluent/plugin/out_forward.rb:494:in `resolved_host'

• fluentd Pod を手動で再起動しても効果はありません。