Open HaveFun83 opened 4 hours ago
@m4us1ne thanks for drill down this issue ;)
looks like https://github.com/torvalds/linux/commit/e92e991748484a8b158cd05e348002d2eeefd1f0 is already in kernel v6.11.6
and newer so this should already be fixed in testing
. Can you test testing
stream 41.20241109.2.0
that was released this week?
Describe the bug
fcos stable
41.20241027.3.0
with kernel 6.11.5 was rolled out today morning on our k8s nodes which broke our CNI (calico). We need to manually rollback to "40.20241019.3.0"We found out that kernel 6.11.5 has a bug which broke the CNI operation
Ref: https://github.com/tailscale/tailscale/issues/13863 https://github.com/torvalds/linux/commit/306ed1728e8438caed30332e1ab46b28c25fe3d8
Reproduction steps
Deploy calico CNI on fcos 41.20241027.3.0 k8s nodes
Expected behavior
calico-node pods can operate without any errors
Actual behavior
calico-nodes pods are unable to operate and never get ready
System details
stable 41.20241027.3.0
Butane or Ignition config
No response
Additional information
looks like the bug was fixed in 6.11.6