Fluentd daemonset crashing

[ctas582]

When I deploy the EFK stack using:

container.training/slides/kube/logs-centralized.md

Line 43 in 7069682

kubectl apply -f https://goo.gl/MUZhE4

the fluentd pods keep crashing with the following error message:

2018-08-13 18:48:12 +0000 [error]: unexpected error error_class=Errno::EACCES error=#<Errno::EACCES: Permission denied @ rb_sysopen - /var/log/fluentd-containers.log.pos>

A recent update to the fluentd-kubernetes-daemonset README says that fluentd needs to be run as the root user, which can be done by setting the FLUENT_UID envvar to "0". Deploying that additional envvar does fix it.

I'm assuming this YAML used to work, but I've been unable to trace what has changed which means it doesn't though. Have other people seen this?

[bridgetkromhout]

Thanks for finding this! We'll take a look and see if we can reproduce.

[jpetazzo]

Ah yes, my buddy @abuisine suggested exactly this (setting FLUENT_UID to 0) and
he linked to this → fluent/fluentd-kubernetes-daemonset#164

Keeping this issue open to track it.