This message was deleted.
# extensions
a
This message was deleted.
s
Glad you managed to resolve the issue. Think this relates to something myself and @wooden-engine-7388 hit when working on changes to
nuxt
. They're only in
master
though, so if you'd like a more stable platform you could switch to the
release-2.7.2
branch
The dashboard communicates with multiple rancher APIs, so as found out the
API
env var should point to the root Rancher url
w
I just wasn't certain, only because Rancher includes the /v3 URL in the UI under the user settings, and with that on top of the connection failure happening. Happy to have it working again
🙌 1
This is really a side note, but I struggled with getting tools installed and synced, since Rancher 2.7.1 supports k8s 1.24, and Rancher Desktop stable is on 1.25. I know the versions wont always line up.
Do you have any guidance on which version of Rancher to deploy for the API server? For staying in sync for local dev
I assumed and used the rancher-latest helm repo
s
I've never really played around with Rancher Desktop (am on linux), is there a way to set the k8s version or is it always fixed? In terms of min/max kube versions for Rancher Manager, they're listed here. Unfortunately we don't publish which k8s versions are supported in un-released rancher manager versions, but 1.25 is currently targetted for 2.7.2 which is due out soon. As usual with latest/dev/head/unreleased builds there may be dragons, but if you wanted to keep your dashboard
master
running with a reasonable Rancher Manager backend you could run the docker image
rancher/rancher:v2.7-head
. That will be bleeding edge though, and not fit for production. For the dashboard
release-2.7.2
branch there should be a
rancher/rancher:v2.7.2-rc<number>
image (replace number with latest RC). In helm world, i'm not sure latest tracks dev
w
Yea it's easy to switch versions in Rancher Desktop
I'm trying to put together a getting started blog and wanted to keep all the tools in the same family
As a part of that, I threw together some more in-depth instructions for rancher management server rather than assuming one is already available
That's great to know about the image, that'll be a lifesaver. Do you by chance already have a skaffold config already thrown together for this or dashboard development?
s
Thinking about it, you might be able to deploy the latest Rancher Manager via helm using something like
Copy code
helm install rancher rancher-latest/rancher --namespace cattle-system --set hostname=... --set rancherImage=rancher/rancher --set rancherImageTag=v2.7head --set rancherImagePullPolicy=Always
again though, that's just for dev and not production. that would unlock 1.25 if you needed it
w
Yea that's exactly what I did, minus the imagetag
s
in terms of deploying Rancher Manager, there's the main docs page that has the best advice. For running the dashboard it should just be a one liner
w
If I put together a skaffold file, I'll have a good all-in-one dev environment as well. I had a little bit of trouble figuring out how to build rancher manager for dev
Was already typing that when you replied with the docs haha
Where this is heading, is I'd like to run the extension kit, rancher manager, and the helm charts for an application we'd be building the extension for, all in one go.
s
ah yeah, we have nothing that brings all of those together
there's some automation we use for e2e tests (see
docker:local:start
target in
package.json
), but that works with docker and has some very weak setup input
We also have CI to build a test extension (at npm dep level), the app (at npm dep level) and run them all together (
/shell/scripts/test-plugins-build.sh
)
w
I can take a stab at it. The makefile approach to the RMS is something I haven't dealt with in a minute, so it blocked me a little bit on the skaffold front
s
We'd love to know how you get on, both with automation of environment and the blog. If you have any other questions feel free to ask