Mirror
/
k3s
mirror of https://github.com/k3s-io/k3s.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

README.md: fix rootless description 

Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
This commit is contained in:
Akihiro Suda 2019-05-29 15:29:18 +09:00
parent 9efbeb11e4
commit 3b8da28eb4
1 changed files with 15 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
README.md
View File

@ -539,13 +539,25 @@ k3s server --node-label foo=bar --node-label hello=world --node-taint key1=value
## Issues w/ Rootless
### Ports
When running rootless a new network namespace is created. This means that k3s instance is running with networking
fairly detached from the host. The only way to access services run in k3s from the host is to setup port forwards
to the k3s network namespace. We have a controller that will automatically bind 6443 and any service port to the
host with an offset of 10000. That means service port 80 will become 10080 on the host. Once you kill k3s and then
start a new instance of k3s it will create a new network namespace, but it doesn't kill the old pods. So you are left
to the k3s network namespace. We have a controller that will automatically bind 6443 and service port below 1024 to the host with an offset of 10000.
That means service port 80 will become 10080 on the host, but 8080 will become 8080 without any offset.
Currently, only `LoadBalancer` services are automatically bound.
### Daemon lifecycle
Once you kill k3s and then start a new instance of k3s it will create a new network namespace, but it doesn't kill the old pods. So you are left
with a fairly broken setup. This is the main issue at the moment, how to deal with the network namespace.
The issue is tracked in https://github.com/rootless-containers/rootlesskit/issues/65
### Cgroups
Cgroups are not supported
## Running w/ Rootless
Just add `--rootless` flag to either server or agent. So run `k3s server --rootless` and then look for the message