New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
v5 bug: Logging to ES Stops and Fluentd cpu -> 100% as ES buffer used #885
Comments
It should be noted that calling the So: |
I have a cluster with identical configuration in production This does does not exhibit the same behavour. So this suggests a new bug introduced in v5. |
Rollback-ed back to v1.11 on one cluster that was misbehaving, and thus far cpu re-away has not been seen. |
What is the version of back to normal? Is it in Fluentd v1.11.5? |
It seems that this CPU usage spike comes from Just changes for supporting IPv6 string on host address. I suspect that this spike is originated from |
And could you send |
Do you intent to getting stacktrace? (I also want to getting it) |
fluent/fluentd#3387 might be a same issue. |
The fluentd I rolled back to is v1.11.1.
|
I tried rolling back just the ES plugin (so fluentd v1.12.3 & ES 4.1.4) this also failed as recorded here fluent/fluentd#3382 This also leads me to suspect that it is a core issue. I will roll forward to v1.12.3/5.0.3 and attempt to get a |
5 node EKS development cluster with very little load Fluentd v1.12.3/ES pluging 5.0.3 fluentd-kqlvw:
fluentd-wrpmk:
|
@andrew-pickin-epi I can see the same endless loop of |
fluent/fluentd#3387 (comment) by @dzikismigol
I found a resolv's issue for it: ruby/resolv#11 I think it's triggered by excon's change: excon/excon#739 after:
before:
|
Not seen this error since moving to 1.13 |
I can report the same, fixed in 1.13 |
Kubernetes AWS/EKS installation, same behaviuour seen on two separate clusters ( albeit with same config).
Problem
Periodically (and with no indication why) Fluentd's ES buffer starts to fill.
CPU consumption by flunetd pod swiftly reaches 100%.
Somethings fluentd is restarts on its own - no human intervention as here:
2021-05-17 13:48:59 +0000 [info]: fluent/log.rb:329:info: Worker 0 finished unexpectedly with signal SIGKILL
and re-connection with ES is established and log begin.
No intervention is performed by users. Config renames unchanged throughout.
Expected Behavior or What you need to ask
...
Using Fluentd and ES plugin versions
fluent-gem list
Pod Env
Plugin config
ES log shows no errors. Other nodes in the daemonset continue to function normally.
Fluentd shows no errors during the period, is not trying to reconnect.
The text was updated successfully, but these errors were encountered: