fleet icon indicating copy to clipboard operation
fleet copied to clipboard

[SURE-8883] Add the options for the user to add custom or automatic tolerations to tolerations for Helm operation pods

Open kkaempf opened this issue 11 months ago • 4 comments

SURE-8883

Motivation

Whenever a user does any operation on a helm chart, a helm-operation-xxx pod is created to execute the operation. Customers that have all nodes tainted will have a problem with this.

Changes were made to the backend SURE-8593 so that a customer may provide tolerations to helm operation pods or ask that it automatically adds the needed tolerations to run on control plane nodes. The changes were made to the install/upgrade/uninstall charts endpoints.

See JIRA for an example of payload

Acceptance criteria

The user can manually specify tolerations to be added to a helm operation pod when installing/upgrading/uninstalling a chart The user can specify that the helm operation pods should add tolerations to all taints from control plane nodes.

kkaempf avatar Feb 07 '25 10:02 kkaempf

Tentatively adding to v2.11 so engineering can triage it and give a rough estimate. The final schedule will be determined based on the estimate (and other backlog items 😉 )

kkaempf avatar Feb 07 '25 10:02 kkaempf

For Fleet:

  • [x] make sure tolerations are passed to the fleet helm chart
  • [ ] make sure fleetcluster passes tolerations to local cluster
  • [x] fleet chart uses tolerations for cleanup/etc. jobs
  • [ ] fleet controller uses tolerations for apply/clone job pods

manno avatar Feb 12 '25 14:02 manno

UI received something similar and has a draft to help out, but we seem to be blocked by either this or https://github.com/rancher/rancher/issues/49101

Dashboard: https://github.com/rancher/dashboard/issues/13125 Internal reference: SURE-9823 (blocked by SURE-8883 mentioned at the top)

@kkaempf as I help out @momesgin with his ticket, does his ticket look connected?

gaktive avatar Feb 20 '25 20:02 gaktive

https://github.com/rancher/rancher/pull/49355 was pushed to rancher/rancher to add tains in control plane as tolerations to the the Fleet's local cluster.

0xavi0 avatar Mar 13 '25 08:03 0xavi0

Done.

kkaempf avatar Aug 14 '25 15:08 kkaempf