How to proceed for Version Upgrades or adding nodes

[noesberger]

Hello

What is the process for upgrading the kubernetes version? Do I just have to change the version in group_vars and then run the playbook again or is this not supported with the current playbook.

Can I extend an installed cluster with more nodes. Do I just have to add more nodes to the inventory and then run the playbook again?

Thanks for your help.

[aslanpour]

I just replaced the version number in group_vars and ran the playbook again to upgrade and it works perfectly.

I had to restart nodes to receive the effect.

In my case, ansible also returned a message that is shown below.

[mvhirsch]

In my case, ansible also returned a message that is shown below.

The error you're facing is already described in #188.

---

As for the upgrade: I think updating the variable and re-run the playbook is fine for now (see also https://github.com/k3s-io/k3s-ansible/issues/94#issuecomment-715667638). Please keep in mind, that any upgrade may break you system. Also check keep in mind that any bugfixes done in scripts should be mirrored on your system, too. For a safe upgrade path I suggest:

1. Drain your node(s)
2. Upgrade your nodes using k3s-ansible, restart them afterwards
3. Drain your master(s) and upgrade them too, restart them afterwards
4. make sure everything is up and running

[RodriMora]

How about addint new nodes? just editing the host.ini file and adding new masters, or nodes works?

And then running the sites.yml again?

[mvhirsch]

Yes, just add them to your hosts.ini and rerun the playbook. I just did that a few weeks ago.

RodriMora @.***> schrieb am Di., 12. Juli 2022, 12:17:

How about addint new nodes? just editing the host.ini file and adding new masters, or nodes works?

— Reply to this email directly, view it on GitHub https://github.com/k3s-io/k3s-ansible/issues/186#issuecomment-1181582563, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAEL456EBGXECD56UID5KWDVTVA3PANCNFSM5UDO6W7A . You are receiving this because you commented.Message ID: @.***>

[badsmoke]

i have now tried to upgrade my cluster via ansible, seems to work for now at first glance

from a 1 month old version with k3s v1.23.4+k3s1 to the current

but at the end of the metalb, an error occurs

Thursday 08 September 2022  13:34:25 +0200 (0:00:01.314)       0:01:48.716 **** 
ok: [10.10.0.10] => (item=controller)
ok: [10.10.0.10] => (item=webhook service)
ok: [10.10.0.10] => (item=pods in replica sets)
failed: [10.10.0.10] (item=ready replicas of controller) => {"ansible_loop_var": "item", "changed": false, "cmd": ["k3s", "kubectl", "wait", "replicaset", "--namespace=metallb-system", "--selector=component=controller,app=metallb", "--for=jsonpath={.status.readyReplicas}=1", "--timeout=60s"], "delta": "0:00:00.187160", "end": "2022-09-08 13:34:32.641412", "item": {"condition": "--for=jsonpath='{.status.readyReplicas}'=1", "description": "ready replicas of controller", "resource": "replicaset", "selector": "component=controller,app=metallb"}, "msg": "non-zero return code", "rc": 1, "start": "2022-09-08 13:34:32.454252", "stderr": "error: readyReplicas is not found", "stderr_lines": ["error: readyReplicas is not found"], "stdout": "replicaset.apps/controller-57fd9c5bb condition met", "stdout_lines": ["replicaset.apps/controller-57fd9c5bb condition met"]}
failed: [10.10.0.10] (item=fully labeled replicas of controller) => {"ansible_loop_var": "item", "changed": false, "cmd": ["k3s", "kubectl", "wait", "replicaset", "--namespace=metallb-system", "--selector=component=controller,app=metallb", "--for=jsonpath={.status.fullyLabeledReplicas}=1", "--timeout=60s"], "delta": "0:00:00.215121", "end": "2022-09-08 13:34:33.171628", "item": {"condition": "--for=jsonpath='{.status.fullyLabeledReplicas}'=1", "description": "fully labeled replicas of controller", "resource": "replicaset", "selector": "component=controller,app=metallb"}, "msg": "non-zero return code", "rc": 1, "start": "2022-09-08 13:34:32.956507", "stderr": "error: fullyLabeledReplicas is not found", "stderr_lines": ["error: fullyLabeledReplicas is not found"], "stdout": "replicaset.apps/controller-57fd9c5bb condition met", "stdout_lines": ["replicaset.apps/controller-57fd9c5bb condition met"]}
failed: [10.10.0.10] (item=available replicas of controller) => {"ansible_loop_var": "item", "changed": false, "cmd": ["k3s", "kubectl", "wait", "replicaset", "--namespace=metallb-system", "--selector=component=controller,app=metallb", "--for=jsonpath={.status.availableReplicas}=1", "--timeout=60s"], "delta": "0:00:00.178730", "end": "2022-09-08 13:34:33.662630", "item": {"condition": "--for=jsonpath='{.status.availableReplicas}'=1", "description": "available replicas of controller", "resource": "replicaset", "selector": "component=controller,app=metallb"}, "msg": "non-zero return code", "rc": 1, "start": "2022-09-08 13:34:33.483900", "stderr": "error: availableReplicas is not found", "stderr_lines": ["error: availableReplicas is not found"], "stdout": "replicaset.apps/controller-57fd9c5bb condition met", "stdout_lines": ["replicaset.apps/controller-57fd9c5bb condition met"]}

[Kybeer]

I just replaced the version number in group_vars and ran the playbook again to upgrade and it works perfectly.

I had to restart nodes to receive the effect."Destination /usr/local/bin/k3s is not writable", "owner": "root", "secontext": "system_u:object_r:bin_t:s0", "size": 53239808,

In my case, ansible also returned a message that is shown below.

I did the same thing - changed K3s version in group_vars/all.yml - but when I re-run the Playbook I get this error on all my nodes:

"Destination /usr/local/bin/k3s is not writable", "owner": "root", "secontext": "system_u:object_r:bin_t:s0", "size": 53239808,

[tstackhouse]

I did this, and jumped right to the latest 1.25 from 1.22. And that resulted in a lot of weirdness that I'm trying to triage over in https://github.com/k3s-io/k3s/issues/6314. Don't do that.

[MalteMagnussen]

I'd also like to know how to upgrade a k3s installed with this Ansible code.

Should be a section in the README describing how to do this imo

[dereknola]

Now Supported via playbook.