Bring back the ability to drain/cordon upstream cluster nodes in 2.6

[andygodish]

Is your feature request related to a problem? Please describe.

In Rancher 2.5, users had the ability to cordon/uncordon & drain upstream cluster nodes. This feature (as it pertains to management of the upstream cluster), was removed from Rancher 2.6. During periods of scheduled maintenance it is convenient to allow our Rancher users to simply drain or cordon a node from the UI. Users of our new 2.6 cluster miss this feature.

Describe the solution you'd like

Add functionality in 2.6 that allows authorized users to perform drain/cordon actions on upstream nodes from the cluster explorer.

Additional context

Image of 2.5 UI allowing management of local cluster nodes:

Feature exists in 2.6, but only for downstream clusters:

No longer able to perform these actions on upstream cluster nodes:

[stale[bot]]

This repository uses a bot to automatically label issues which have not had any activity (commit/comment/label) for 60 days. This helps us manage the community issues better. If the issue is still relevant, please add a comment to the issue so the bot can remove the label and we know it is still valid. If it is no longer relevant (or possibly fixed in the latest release), the bot will automatically close the issue in 14 days. Thank you for your contributions.

[felixng-umh]

same here

[github-actions[bot]]

This repository uses an automated workflow to automatically label issues which have not had any activity (commit/comment/label) for 60 days. This helps us manage the community issues better. If the issue is still relevant, please add a comment to the issue so the workflow can remove the label and we know it is still valid. If it is no longer relevant (or possibly fixed in the latest release), the workflow will automatically close the issue in 14 days. Thank you for your contributions.

[govbetrieb]

I also got this problem

[richard-cox]

From the UI's perspective in 2.6.6 the actions are in place (cordon, uncordon, drain, stopdrain) but are gated on whether the v3/node resource has the matching links in it's action collection. Looks like for these cases the local cluster's nodes does not have them.

To assist investigating this, please could commenters add the type of kube Rancher is deployed in and the roles the nodes have (where cordon, etc is expect to show)?

[bashofmann]

In my test setup, where I can reproduce it: Rancher 2.6.6 on RKE2 1.21.9. All nodes have all 3 roles.

Interestingly the cluster is detected as a k3s cluster on the cluster detail page

[govbetrieb]

RKE provider

[richard-cox]

@govbetrieb In your browser can you go to <rancher url>/v3/nodes, find the entry for k8s-worker04 and check that the actions property contains entries such as cordon, uncordon, drain, stopdrain?

[govbetrieb]

I hope i got this right, here you are, name of the Node is on the bottom line

[richard-cox]

@govbetrieb Perfect, thank you. This is a UI issue so i'll move over to our repo

[Josh-Diamond]

Ticket #6641 - Test Results

With HA Jenkins:

Reproduced on rancher v2.6.6:

1. Fresh install of rancher v2.6.6
2. As admin, navigate to Cluster Explorer > Nodes of the local cluster
3. Reproduced - No cordon/drain options

Screenshot:

---

Verified on rancher v2.6-4f2da17be6d673ce6dba831c2dd560d3b3cb2ed1-head:

1. Fresh install of rancher v2.6-head
2. As admin, navigate to Cluster Explorer > Nodes of the local cluster
3. Verified - Cordon/Drain options

Screenshot:

[Josh-Diamond]

Re-opening for further validation

[Josh-Diamond]

TIcket #6641 - Test Results [pt. 2]

With HA Helm + rke1 local - [3 nodes; all-roles]:

Verified on rancher v2.6-633f5aadaf0768f85f7ada74888f8392d374a394-head:

1. Ensure local cluster kube node's can determine their cluster id
2. Verified - Nodes listed in Cluster Explorer of local; Nodes listed under machine pools of Cluster Management; accurate provider in Cluster Dashboard
3. Ensure cordon/drain functionality works w/ RKE1 local
4. Verified - cordon/drain available and functional for nodes with worker roles - [including all-roles]

---

With HA Helm + rke1 local - [3 nodes; 1 etcd, 1 cp, 1 wkr]:

Upgrade verified on rancher v2.6.7 => v2.6-head:

1. Fresh install of rancher v2.6.7
2. Ensure cordon/drain functionality missing w/ RKE1 local; Cluster Explorer > Nodes
3. Verified - No cordon/drain functionality or options available
4. Upgrade rancher to v2.6-head
5. Ensure cordon/drain options are seen + functional for nodes w/ worker roles
6. Verified - cordon/drain available and functional for nodes w/ worker roles

---

With HA Helm + rke2 local - [4 nodes; 3 etcd/cp, 1 wkr]:

Verified on rancher v2.6-633f5aadaf0768f85f7ada74888f8392d374a394-head:

1. Fresh install of rancher v2.6-head
2. Ensure cordon/drain functionality works w/ RKE2 local
3. Verified - cordon/drain available and functional for nodes w/ worker roles
4. Ensure local cluster kube node's can determine their cluster id
5. Verified - Nodes listed in Cluster Explorer of local; Nodes listed under machine pools of Cluster Management; accurate provider in Cluster Dashboard