canonical / microk8s.io Goto Github PK

should read

alias kubectl='microk8s.kubectl'

https://microk8s.io/docs/ports#authentication-and-authorization

"Upon deployment MicroK8s creates a CA, a signed server certificate and a service account key file. These files are stored under /var/microk8s/current/certs."

I think the location is (now) incorrect:

ls -l var/microk8s/current/certs
ls: cannot access 'var/microk8s/current/certs': No such file or directory

Should be:

ls -l /var/snap/microk8s/current/certs
total 48
-rw-r--r-- 1 root root 1115 Jun 22 18:59 ca.crt
-rw------- 1 root root 1675 Jun 22 18:59 ca.key
-rw-r--r-- 1 root root   41 Jun 22 18:59 ca.srl
-rw-r--r-- 1 root root  695 Jun 22 18:59 csr.conf
-rw-r--r-- 1 root root  695 Jun 22 19:25 csr.conf.rendered
-rw-r--r-- 1 root root  665 Jun 22 18:59 csr.conf.template
-rw-r--r-- 1 root root 2287 Jun 22 18:59 kubelet.crt
-rw------- 1 root root 1675 Jun 22 18:59 kubelet.key
-rw-r--r-- 1 root root 1537 Jun 22 18:59 server.crt
-rw-r--r-- 1 root root 1261 Jun 22 18:59 server.csr
-rw------- 1 root root 1679 Jun 22 18:59 server.key
-rw------- 1 root root 1675 Jun 22 18:59 serviceaccount.key

Write deploy-to-staging.microk8s.io & deploy-to-microk8s.io Jenkins jobs, and run them to deploy the websites to Kubernetes.

Due: Nov 18th (Today)

Note; this appears on both ubuntu.com and microk8s.com

"Need to find a way to tell people visiting Microk8s.io during Kubecon that they can come and visit the Canonical / Ubuntu booth to win a MicroK8s t-shirt."

Brief

Requested by Thibaut.

There are some text links in the docs wich should be converted.

The Kubernetes image in the hero area is not high definition for retina screens.

https://microk8s.io/docs/

links to https://raw.githubusercontent.com/cncf/artwork/master/kubernetes/certified-kubernetes/versionless/color/certified-kubernetes-color.png which is broken.

Firefox 66.0.3, on Ubuntu 18.10 Cosmic Cuttlefish

from https://github.com/ubuntu/microk8s/blob/master/docs/registry.md

Highlight the page you are on.

Instead of /var/snap/microk8s/current/credentials/basic_token.csv now it's /var/snap/microk8s/current/credentials/basic_auth.csv

We need to add the Canonical global navbar to the microK8s site for consistency with the rest of our websites.

The page "Working with image registries" (https://microk8s.io/docs/working) has messed up formatting because the example error message is toooooo looooong.

This (unnecessarily) makes the whole page difficult to read.

The microk8s.io homepage has a "Get it from the Snap Store" badge, but instead of hot-linking the image (as it should) it uses a copy from assets.ubuntu.com

https://microk8s.io/

Expected: hotlinks from https://snapcraft.io/static/images/badges/en/snap-store-white.svg
Actual: loads from https://assets.ubuntu.com/v1/f0bd37f9-snap-store-white-231118.svg

See Peter's comment ↓

It's too big!

Improve the way the ticket list looks on microk8s homepage.

how it currently looks:

how it should look like:

from https://github.com/ubuntu/microk8s/blob/master/docs/release-channels.md

The docs for dashboard addon do not work.
I get to the login screen, select the "Token" option to login and paste the long token string that was generated by the commands described in that guide. I click the signin button and nothing happens.

I now stop evaluating microk8s - this seems to be in a very early stage. I really do not understand why you are advertising this at the login screens of server installations - it gives a very bad impression. I would not write that if it was not Ubuntu. My god, what a bad experience, I can not believe it, this is Ubuntu!!! - you can not do that, it is awful, it will destroy everything. Stop it!

Write deployment configs for microk8s.io in https://github.com/canonical-webteam/deployment-configs/

As seen in https://twitter.com/ktsakalozos/status/1049568237296513024, https://microk8s.io/assets/img/card.jpg directs readers to install the snap from the beta channel, whereas the main site omits a channel (therefore, picks stable)

What's the issue:
In recent version of the docs the whole section about adding alias has been removed.

Expected:
In documentation for our project https://mojaloop.io/documentation/deployment-guide/ we rely on a fact that section about aliast is there.

Proposed fix:
Add back the missing section.
Or reject that ticket, which means we need to adjust documentation in our project (and probably some other dependent project will need that too).

Details:
Link to old version: https://github.com/canonical-web-and-design/microk8s.io/blame/7aeaad66ecd8d961f9527b7328cfae22398ccbdf/docs/index.md
Link to new version: https://github.com/canonical-web-and-design/microk8s.io/blame/master/docs/index.md

I enabled ingress controller . I seen its ngjnx based . For testing one of my service its need more than 1gb need to uploaded. But in this setup its failed because of 413 large entity problem. Any workaround for this ?

Should we redirect traffic going to microk8s.com to microk8s.io? Is this something we do or we just say ".io is the only right location"?

On the main page (https://microk8s.io/) there is snap info microK8s (capital letter).
It does not work after copying, pasting and executing:

[root@centos ~]# snap info microK8s
error: no snap found for "microK8s"

Should be: snap info microk8s.

Please update index.html - "microk8s - Zero-ops upstream Kubernetes in a cross-distro single-package" from the copy doc

There is an error with this repository's Renovate configuration that needs to be fixed. As a precaution, Renovate will stop PRs until it is resolved.

Error type: undefined

Add our standard structure to the repository, so it can run with ./run, and a Docker image can be built for deploying to production. Also ass /_status/check endpoint.

Followed the Add Node tutorial up to point where I attempted to join the node. There was no confirmation or anything. I then attempted to run microk8s.kubectl get no and got the message in the title.

I copied and pasted the commands and results below.

Running MicroK8s for Ubuntu 18.04 on a Dell G5.

$ microk8s.add-node
Join node with: microk8s.join 10.56.253.71:25000/fWfNqtbAmINQNgFYdBjlelYPjKpscWQp

If the node you are adding is not reachable through the default interface you can use one of the following:
microk8s.join 10.56.253.71:25000/fWfNqtbAmINQNgFYdBjlelYPjKpscWQp
microk8s.join 192.168.39.1:25000/fWfNqtbAmINQNgFYdBjlelYPjKpscWQp
microk8s.join 192.168.122.1:25000/fWfNqtbAmINQNgFYdBjlelYPjKpscWQp
microk8s.join 172.17.0.1:25000/fWfNqtbAmINQNgFYdBjlelYPjKpscWQp
microk8s.join 10.1.99.0:25000/fWfNqtbAmINQNgFYdBjlelYPjKpscWQp

$ microk8s.join 10.56.253.71:25000/fWfNqtbAmINQNgFYdBjlelYPjKpscWQp
$ microk8s.kubectl get no
This MicroK8s deployment is acting as a node in a cluster. Please use the microk8s.kubectl on the master.

It's worth noting that that the installation is kind of wonky. Attempting to start ubuntu normally freezes after login. I've been working around it by Going to advance boot, selecting to start in recovery mode, then selecting to boot normally. Not sure if this would have any impact on the problem reported.

NB: Only do this after #8

Ask IS to point DNS for staging.microk8s.io and microk8s.io at prod-comms.kubernetes-ingress.canonical.com.

@evilnick, a user asked on #microk8s, how/why an upgrade broke his deployment:

Here are some good practices you should consider in order to have a more stable
setup.

• Do not follow the latest channel. The latest channel will get major upgrades when upstream Kubernetes releases.
  For example the 1.16 release (next week) will transition all MicroK8s instances to the flannel cni and introduce clustering.
  When stability is important you should follow a specific channel with for example sudo snap install microk8s --classic --channel=1.15/stable.
  Please have a look at the release scheme we follow at https://microk8s.io/docs/release-channels.
• Snap upgrades happen transparently in the background but you can schedule them in a convenient to you time.
  You can do that with snap set system refresh.timer as shown in https://snapcraft.io/docs/keeping-snaps-up-to-date
  You can introduce a delay to upgrades of critical sections of your infrastructure so as to first test the upgrade process in a non critical sections.
• To have full control on what revisions land and when you could consider setting up a proxy https://docs.ubuntu.com/snap-store-proxy/en/
• Please subscribe to the microk8s tag at https://discuss.kubernetes.io/tags/microk8s so you get announcements.
• We obviously do not want to break people during upgrades. So we run upgrade tests before every release.
  If you are interested I could show you how the upgrade is tested. Should we find that we have gaps
  important to you, we have to close them so we do not break you in the future.
  Here is the upgrade test we have https://github.com/ubuntu/microk8s/blob/master/tests/test-upgrade.py
• In the future, if you see some unexpected behavior please run a microk8s.inspect and attach the produced tarball on an issue at
  https://github.com/ubuntu/microk8s/issues .

I have followed the instructions for using the dashboard add-on; however, i am not able to access the dashboard following the instructions provided. i have selected the non-rbac option, copied my token, enabled the proxy via an ssh terminal and copied the dashboard url into a browser. With the loopback address 127.0.0.1 i do not connect to the dashboard. If i change the loopback IP address to the host IP address, i get the dashboard authentication process. I paste the token into the web field, select the signin button, but nothing happens. The dashboard remains in the authentication webpage.

https://microk8s.io/docs/addon-dashboard

At the top the site reads:
"A single package of k8s that installs on 41 (stroked out) 42 flavours of Linux. Made for developers and great for appliances."

Using the copy doc and design to build a demo

based on https://github.com/ubuntu/microk8s/blob/master/README.md

The link in the following gives a 404 error

you go though the list of ports that need to be available

Docs for installing on macos (specifically multipass) are inconsistent. The multipass section here tells us to use at least 4g ram / 40g disk, multipass exec to install microk8s and configure the network, but nothing about enabling services:

https://microk8s.io/docs/install-alternatives

Compare that to the macOS section of the getting started page, which does not tell us to use any minimums, nor exec/shell to install or configure the network, but it does enable services:

https://microk8s.io/#get-started

These should be synchronized so users come out the other side with a recommended environment.

move https://github.com/ubuntu/microk8s/blob/master/docs/dockerd.md

Currently, all the assets that make up the asciinema video adds up to over 1MB in size. I think we should convert it into a short video showing the same thing.

The following commands mentioned in the documentation is missing from the snap (find details below).

• microk8s.status
• microk8s.stop
• microk8s.start

Am I missing any steps ? First time using snap so could be I have screwed up something in the setup.

I cant find these commands in the manifest.yaml as well.

snap info microk8s
name:      microk8s
summary:   Kubernetes in a snap.
publisher: Canonical*
license:   unset
description: |
  MicroK8s deploys and configures all kubernetes services in a snap.
commands:
  - microk8s.config
  - microk8s.disable
  - microk8s.docker
  - microk8s.enable
  - microk8s.inspect
  - microk8s.istioctl
  - microk8s.kubectl
  - microk8s.reset
services:
  microk8s.daemon-apiserver:          simple, enabled, active
  microk8s.daemon-controller-manager: simple, enabled, active
  microk8s.daemon-docker:             simple, enabled, active
  microk8s.daemon-etcd:               simple, enabled, active
  microk8s.daemon-kubelet:            simple, enabled, active
  microk8s.daemon-proxy:              simple, enabled, active
  microk8s.daemon-scheduler:          simple, enabled, active
snap-id:      EaXqgt1lyCaxKaQCU349mlodBkDCXRcg
tracking:     stable
refresh-date: yesterday at 13:51 EST
channels:                                          
  stable:         v1.12.2  (266) 224MB classic <
  candidate:      v1.12.2  (266) 224MB classic 
  beta:           v1.12.2  (266) 224MB classic 
  edge:           v1.12.2  (313) 224MB classic 
  1.12/stable:    v1.12.2  (267) 224MB classic 
  1.12/candidate: v1.12.2  (267) 224MB classic 
  1.12/beta:      v1.12.2  (267) 224MB classic 
  1.12/edge:      v1.12.2  (314) 224MB classic 
  1.11/stable:    v1.11.4  (255) 219MB classic 
  1.11/candidate: v1.11.4  (255) 219MB classic 
  1.11/beta:      v1.11.4  (255) 219MB classic 
  1.11/edge:      v1.11.4  (292) 219MB classic 
  1.10/stable:    v1.10.10 (285) 175MB classic 
  1.10/candidate: v1.10.10 (285) 175MB classic 
  1.10/beta:      v1.10.10 (285) 175MB classic 
  1.10/edge:      v1.10.10 (291) 175MB classic 
installed:        v1.12.2  (266) 224MB classic 

I installed Traefik in microk8s and able to access the Traefik dashboard and see all the ingresses frontends and backends but all of them gives me “Gateway timeout” error. I setup the exact same thing in minikube and it works fine. I installed Traefik into kube-system namespace with correct ClusterRole and ClusterRoleBinding. Is there something missing in microk8s that’s preventing Traefik to work? Any help is much appreciated. Thanks.

Ask IS to create HTTPS certs for staging.microk8s.io and microk8s.io and add them to Kubernetes as secrets.

In the index.md there is a mention about ${SNAP_DATA}/args/kube-apiserver configuration but in the actual microk8s release 1.14.1 rev 522 the folder args doesn't exist. Moreover there are two places where the same configurations are present:

• /snap/microk8s/current/default-args
• /snap/microk8s/current/microk8s-resources/default-args

What is the effective args folder ?

Complete homepage copy refresh.

copy doc

Minor updates requested by SFabel in the copy doc

We should add the following to docs when there is a standard profile:

LXD/LXC

MicroK8s can be run inside an LXC container. The only caveat here is that the
container requires certain settings which are not part of the default LXD
profile.

Two profiles are available in the [MicroK8s repository][profile] - one for
LXD using ZFS storage, and one for the default storage. Use one of
these files to create a new profile:

lxc profile create mk8s
lxc profile edit mk8s < microk8s.profile

... and then run a container with:

lxc launch -p default -p mk8s ubuntu:18.04 microk8s-container

You can then launch a shell in the container and perform commands as
usual:

lxc exec microk8s-container shell

You can also obtain the IP address of the container by running:

lxc list

this can be used to access the API.

Please update index.html - "microk8s - Zero-ops Kubernetes for workstations and edge / IoT" from the copy doc