-
Notifications
You must be signed in to change notification settings - Fork 74
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update or add current Tailscale terminology #439
Comments
Thanks for the feedback, you are right, things has changed. Adding the missing link is a no-brainer. Originally both the reverse proxy and the funnel was an option for the serve command. Additionally the current 2 separate bool config option (proxy and funnel) is also, hmmm, not so good. See also #407. So I suggest to merge them, but I really struggle with a good naming (it's harder than developing this stuff). What do you think about this: config.yaml: - funnel: bool?
- proxy: bool?
+ share: list(disabled|tailnet|tailnet and internet)?
- proxy_and_funnel_port: match(^(443|8443|10000)$)?
+ share_port: match(^(443|8443|10000)$)? And the documentation can explain/reference serve for the tailnet option and funnel for the internet option. Tailscale uses similar wording, like (this is a funnel log):
And uses similar wording in the online serve/funnel documentation "You can think of this as sharing the service, ... with the rest of your tailnet." "You can use it to share a local service," Yeah, changing an existing config is not trivial. At the beginning of stage2_hook.sh we can convert the old config with |
Hi, I am also not good at naming and organizing options. My suggestion for something like this, where you are providing a configuration interface to a separate tool that someone else has developed, is to copy them. They spend the time and effort to figure out how to name and organize things, so you don't have to. :-) One benefit of copying is that you have to write less documentation--just name your option the same, and users can read the other tool's documentation to learn more. I believe that the separate configuration options are still a good idea. Tailscale themselves split them up. So as a general Tailscale user (ignoring HA), I have 3 options.
If I'm using the HA Tailscale Add-On, I want to see separate options to configure 1 & 2. In the documentation, maybe make a mention of 3 because it's a combo of something I configure in the Add-On + something I do on the Tailscale layer). In my case, I wanted to use the One other note, I do like that the docs link to the Tailscale HTTPS docs. IMO, it's important that users understand that enabling Regards, |
Yes, right, the tailnet and internet explanation is already in the docs, I go with |
I admit, I have never configured Have you tried configuring funnel without serve (separate from the Add-On, I mean)? |
Yeah, "independent"... Eg. serve reset will reset funnel also, and vice versa, you serve 2 things then funnel a 3rd, all 3 will be available on the internet (that's why in #407 a second nginx reverse proxy was needed), etc... Instead of independence, they've created a total mess (in my opinion). I've tested the "independence" a few month ago, and I haven't seen anything in the changelog that it has changed. Serve is a "full blown" reverse proxy, and funnel is only a TCP forwarder in front of it that can be enabled or disabled, all config happens inside the reverse proxy (where the https traffic ended). Now the only thing I need is some free time to implement it... |
That may have been true in the past. But it isn't true at all, as of today. I don't know what you're looking for in changelogs. I have one machine currently with a
This is also not true, as evidenced by the docs and my testing. I set up a So please, don't do whatever weirdness you're thinking of doing to the settings. Just copy the settings the way that Tailscale does them. That will make the most sense to users. |
Who is talking about 2 different servers??? Execute (in the add-on):
And you can access both test1 and test2 at https://xxxx.tailxxxx.ts.net:8443/testX So funnel turns on the TCP forwarder, and everything is published on the internet, all serve endponts, not just funnel endpoints. Until TS implements filtering in the reverse proxy, the TCP forwarder will publish everything behind a certain port. And don't forget to reverse the above changes, execute:
And miraculously funel also reseted. Case closed, they are not independent. |
Problem/Motivation
Similar to home-assistant/core#132976, it would probably be a good idea to align the Add-On's terminology, especially documentation, with that of Tailscale.
Expected behavior
tailscale/DOCS.md to mention somewhere that "proxy" = Tailscale serve.
And maybe for the "proxy" YAML item to be named "serve".
Actual behavior
The word "serve" is nowhere on the page, nor is there a link to the Tailscale serve documentation. It links to the HTTPS stuff, and that's good.
Proposed changes
Either rename the "proxy" option to "serve", add the word "serve" or an explanation that "proxy" = "serve" somewhere.
Also, especially add a link to the Tailscale serve documentation. https://tailscale.com/kb/1312/serve
I understand that for novice users, the term "proxy" may be desirable because that's what Tailscale serve does. I also recognize that I don't know how easy it would be to change the "proxy" item in the YAML config. So I'm not going to suggest that those get removed entirely. But for those of us who know something about Tailscale, mentioning "serve" somewhere would be helpful.
For me, it was only the repeated alignment of "proxy" with "funnel", and the vague recollection I have that "serve" might have had another name in the past that made me try it out. The logs confirmed it was really Tailscale serve.
The text was updated successfully, but these errors were encountered: