Add support to run with podman in rootless mode using cgroups v2

[gonzalesraul]

What

Fix running k3d with podman and cgroups (v2)

Why

• By running podman in rootless mode, k3d/k3s requires extra delegation in order to work (cpu, cpuset and I/O delegations), added the steps to the doc. Some errors like failed to find cpu cgroup (v2) would happened on cgroup v2 and cpu delegation not enabled ([BUG] rootless docker -> k3d blocks forever (k3s boot loops) #585)

• Podman is API compatible with docker, however there is a slightly difference on network retrieval of IP address. One uses only CIDR notation (127.0.0.1/32) while the other, for unique IP address use the literal notation (127.0.0.1). Create a wrapper function to extract the value instead of throwing an error.

  The following happened before the fix (netaddr.ParseIPPrefix("11.89.1.4"): no '/' )

  ❯ k3d cluster create testing
  INFO[0000] Prep: Network                                
  INFO[0000] Created network 'k3d-testing'                
  INFO[0000] Created image volume k3d-testing-images      
  INFO[0000] Starting new tools node...                   
  INFO[0000] Starting Node 'k3d-testing-tools'            
  INFO[0001] Creating node 'k3d-testing-server-0'         
  INFO[0001] Creating LoadBalancer 'k3d-testing-serverlb' 
  INFO[0001] Using the k3d-tools node to gather environment information 
  INFO[0002] HostIP: using network gateway 10.89.1.1 address 
  INFO[0002] Starting cluster 'testing'                   
  INFO[0002] Starting servers...                          
  INFO[0002] Starting Node 'k3d-testing-server-0'         
  INFO[0007] All agents already running.                  
  INFO[0007] Starting helpers...                          
  INFO[0007] Starting Node 'k3d-testing-serverlb'         
  ERRO[0015] Failed Cluster Start: error during post-start cluster preparation: failed to get cluster network k3d-testing to inject host records into CoreDNS: failed to parse IP of container k3d-testing: netaddr.ParseIPPrefix("11.89.1.4"): no '/' 
  ERRO[0015] Failed to create cluster >>> Rolling Back    
  INFO[0015] Deleting cluster 'testing'                   
  INFO[0016] Deleting cluster network 'k3d-testing'       
  INFO[0016] Deleting 3 attached volumes...               
  WARN[0016] Failed to delete volume 'k3d-testing-images' of cluster 'testing': docker failed to delete volume 'k3d-k3s-default-images': Error response from daemon: volume k3d-k3s-default-images is being used by the following container(s): 5fc21a07d4189a0a93dd3a65142f1f5889e1788ca0076347e4d80aed05e9dd9d, 714544cc5de7ae3ae273f6c3c1285099d8509e4a8c3fe27d7c0ac59ff38537de: volume is being used -> Try to delete it manually 
  WARN[0016] Failed to delete volume 'k3d-testing-images' of cluster 'testing': failed to find volume 'k3d-testing-images': Error: No such volume: k3d-testing-images -> Try to delete it manually 
  FATA[0016] Cluster creation FAILED, all changes have been rolled back! 

Implications

N/A

[gonzalesraul]

@all-contributors please add @gonzalesraul for code and doc

[allcontributors]

@gonzalesraul

I've put up a pull request to add @gonzalesraul! 🎉

[iwilltry42]

LGTM!

[iwilltry42]

Hi @gonzalesraul , thanks a lot for this, this is great! 👍