- Provides information explaining that K3s supports helm v2 and v3 as of our v1.17.0 release.
- Links to official documentation which has good instructions on how to leverage helm v2 / v3 and migrate.
- PostgreSQL/MySQL already have links to docs for DSN that users could reference to escape special chars
- Make suggested command work out of the box (won't run into an issue with unescaped dash in db name any more)
- Most contributors won't know to always update this date
- If we actually want to keep this date (which is rendered on the website) we need to ensure all contributors always update the date...
- Initial commit
- Adds this new page describing the flannel backend options and how to set up your own CNI
- Tweaks the existing Networking page
- Adds hyperlinks linking the two pages together.
Note: Intent is that the main Networking page is for CoreDNS, Traefik, ServiceLB, and in the future other general high level networking discussions. Whereas the "network-options" page under the Installation section is for specific things that should be considered at install-time for K3s such as the flannel backend or other CNIs.
- Make it more clear that you can still modify / add / remove labels and taints, just not with k3s after node registration
You can still change these with the kubectl client, so the wording was changed to make this more clear to avoid confusion.
- Also change the Networking page to instead link to the Networking seciton of the NOde Requirements page so we don't duplicate information in two areas.
- Based on discussion with the K3s team we decided to remove these completely
This is because everything prior to v1.0.0 is not GA and not officially supported. It's strongly recommended to use the v1.0.0 or newer version of K3s.
- Modify the Traefik sectoin so it's clear it's the ingress controller
- Make it clear what ports are use don the host (so can't use for HostPort or a strange NodePort setting)
- Small misc tweaks
- As discussed via slack with Craig, completely remove cli flags from this page
-- This content was not helpful. It's the same as running the --help flag we didn't provide any context or additional info.
-- At a later date we may cherry pick some of those cli flags and add more detail about them where necessary.
- Remove the term "master" from everywhere in the docs. k3s servers are servers not "masters"
- Remove an uneeded section from the ha install page
- Misc small tweaks
- Addresses comments in the PR adding to keep this content but move it.
- Added some context to this page and info about how v1.0.0 and newer has the metrics-server by default so this is clear to the user.
- Add an older-installations page, a child of the Installation main section to address older k3s versions different method for bringing up servers
- Modify the main HA install page so that we no longer list anything as experimental
-- Also update the page so we remove the Cluster Datastore Options section, compact this into a modified section of step 2 for bring up the master (server) nodes
-- Fix any wording, modify text so it's more readable and easier to follow.
- We now includ our own metrics-server and thus standard kubectl top nodes, etc functionality works out of the box.
Since this works like standard k8s now with the top command, etc we don't need to document the custom-method to deploy metrics-server so we should simply remove this section.
Claudia Beresford
Catherine Luse
Catherine Luse
Ruben Klink
Marcus Geißler
galal-hussein
David Nuzik
David Nuzik
Vicken Simonian
YANO Tetsuro
David Zisky
Erik Wilson
Craig Jellick