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

TX Packets dropped on flannel.1 #1009

Open
discostur opened this Issue Jun 22, 2018 · 1 comment

Comments

Projects
None yet
2 participants
@discostur

discostur commented Jun 22, 2018

Hi,

we are running a 14 node k8s cluster with flannel (backend: vxlan) and we are seeing some packet drops on the flannel.1 interface:

        TX packets 416348507  bytes 685679045151 (638.5 GiB)
        TX errors 0  dropped 402 overruns 0  carrier 0  collisions 0

Under normal load (cpu about 20%) we are seeing 0.01pps getting dropped. If we are running some high traffic tests with iperf the dropping rate increases up to 1pps.

We are seeing this behaviour also on smaller clusters with only 3 oder 5 nodes.

Your Environment

  • Flannel version: 0.10.0
  • Backend used (e.g. vxlan or udp): vxlan
  • Etcd version: 3.2.15
  • Kubernetes version (if used): 1.10.2
  • Operating System and version: centos7.5
@andyzasl

This comment has been minimized.

andyzasl commented Oct 26, 2018

Have same issue on debian 9 with kernel 4.9.110-3+deb9u5

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment