-
Expected BehaviorProceed with the playbook Current BehaviorWhen running the playbook, it hangs at the I tried a bunch of times to I checked the discussion here but it did not fix the issue I have. if I use my master node ip address like suggested by FrostyFitz in the discussion it will work but if I put another address for the endpoint it does not work. I tried to use either same network for the virtual ip and ip range (10.193.1.1/24) or a different network (10.193.20.1/24) to have more ips but the result is the same. Steps to Reproduce
Context (variables)Operating system: Hardware: Variables Used
k3s_version: v1.25.12+k3s1
ansible_user: ubuntu
systemd_dir: /etc/systemd/system
flannel_iface: "eth0"
apiserver_endpoint: "10.193.20.10"
k3s_token: "sKcyohCecVULptzpvatzHrYagPGL4mfN"
extra_args: >-
--flannel-iface={{ flannel_iface }}
--node-ip={{ k3s_node_ip }}
extra_server_args: >-
{{ extra_args }}
{{ '--node-taint node-role.kubernetes.io/master=true:NoSchedule' if k3s_master_taint else '' }}
--tls-san {{ apiserver_endpoint }}
--disable servicelb
--disable traefik
extra_agent_args: >-
{{ extra_args }}
kube_vip_tag_version: "v0.5.12"
metal_lb_speaker_tag_version: "v0.13.9"
metal_lb_controller_tag_version: "v0.13.9"
metal_lb_ip_range: "10.193.20.20-10.193.20.99"Hosts
[master]
10.193.1.[155:157]
[node]
10.193.1.[158:159]
[k3s_cluster:children]
master
node
LogsOn the master node On the worker node Possible Solution
|
Beta Was this translation helpful? Give feedback.
Answered by
AshDevFr
Sep 26, 2023
Replies: 1 comment
-
|
Ok so after two days testing that here are my findings and what "worked" for me. First problemI was kinda bothered by this: I did not understand why I had 3 DNS (actually 4 in the file) and most of all why I had 2 times each DNS record. It turns out that because I had 2 network interfaces (I wanted to create an internal network for k8s) it was duplicating the DNS records. I removed an interface and it allowed me to use a virtual ip. Not sure it's related to the DNS though. Maybe kube-vip struggles if there are two interfaces? No idea. Second problemI wanted to use a different subnet for the virtual ip and the metallb range. I ended up moving all of the nodes to the new subnet to only have one subnet for all k8s matters. Third problemDNS issues. This fixed the issue. It's not related to this repo at all but it could help others using ConclusionIt seems that a bunch of small issues (some with cloud-init) were preventing this playbook to run. After untangling them and fixing them one by one, it worked. |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
AshDevFr
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Ok so after two days testing that here are my findings and what "worked" for me.
First problem
I was kinda bothered by this:
I did not understand why I had 3 DNS (actually 4 in the file) and most of all why I had 2 times each DNS record. It turns out that because I had 2 network interfaces (I wanted to create an internal network for k8s) it was duplicating the DNS records.
I removed an interface and it allowed me to use a virtual ip. Not sure it's related to the DNS though. Maybe kube-vip struggles if there…