Lars Ekman
Lars Ekman
/cc @saschagrunert
That *is* weird. I use docker which I suppose my stake-holders also do, but I think they are using `cri-o` so they will probably bump into this problem eventually.
@saschagrunert I have not used Podman and I use a local private docker-registry and transfer my local built images to it using `skopeo`, so it may take a while. If...
@saschagrunert Scratch that. I have tried and I get the same problem with Podman. Use Ubuntu 22.04 and podman installed with `apt install`; ``` > podman version WARN[0000] Failed to...
Yes, without "runAsUser: 1000" it works with the docker and the podman built images.
I had similar problems on Ubuntu 20.04, https://bugs.launchpad.net/ubuntu/eoan/+source/iproute2/+bug/1856045. I don't know if it's related. I had to use the setuid-bit instead which is less secure by far.
I run into this when trying antrea with cri-o and I found an "interesting" work-around (beside re-start cri-o): when `/etc/cni/net.d/10-antrea.conflist` appears, wait a few seconds then do; ``` touch /etc/cni/net.d/99-dummy.conf...
My silly work-around start script; ```bash #! /bin/sh if test "$1" != "cni-conf-wait"; then $0 cni-conf-wait > /dev/null 2>&1 & exit 0 fi while ! test -r /etc/cni/net.d/10-antrea.conflist; do sleep...
> If you are just trying to run some pods and not going to use any BGP related functionality of kube-router, it seems specifying router id is unintutive and should...
Ref; https://github.com/cilium/cilium/issues/13019