Closed kappa8219 closed 1 year ago
It is not really an issue, more like a FR :)
Trying cdebug I discovered that containerd is namespaced. And default NS is (surprise) "default". I see no posibillity to change it with parameters.
Thus we have k8s workload containers out of reach:
# ctr -n k8s.io containers list | head -n 2 CONTAINER IMAGE RUNTIME 03b775021b8b6010bdd47c989264ed6e1dc055247bcb4422ecec330fc56f98a1 public.ecr.aws/isovalent/cilium:v1.11.8-eksa.1 io.containerd.runc.v2 # ctr -n default containers list CONTAINER IMAGE RUNTIME cdebug-a442c538 docker.io/library/busybox:latest io.containerd.runc.v2 nginx-1 docker.io/library/nginx:latest io.containerd.runc.v2
Can you please either add containerd namespace as parameter? I'v seen k8s support upcoming feature. Will it be k8s:// prefix as a namespace setter or pods support (like crictl capable). Any of these are nice and expected features.
https://github.com/iximiuz/cdebug/commit/4a4c24157c0f2dbd9c3f720e6833e2dedeca5673#r92063972
Great catch! Fixed in v0.0.5. Here is an example of how to debug a containerd-cri container.
v0.0.5
It is not really an issue, more like a FR :)
Trying cdebug I discovered that containerd is namespaced. And default NS is (surprise) "default". I see no posibillity to change it with parameters.
Thus we have k8s workload containers out of reach:
Can you please either add containerd namespace as parameter? I'v seen k8s support upcoming feature. Will it be k8s:// prefix as a namespace setter or pods support (like crictl capable). Any of these are nice and expected features.