This message was deleted.
# rke2
a
This message was deleted.
👍 1
a
I didn't get to opening a ticket, because shortly after this I identified it was tied to net.ipv4.xfrm4_gc_thresh being too low - increase the number, and the pods that were having issues connecting out to another endpoint suddenly started working
we got the error message 'No buffer space available' when trying to connect to remote services over the ipsec network, but it worked over the host network just fine - cilium was reporting the health endpoint on the worker node was healthy, but the overlay network target was unhealthy over the ipsec network
c
interesting. thanks for the follow-up!