Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
Repo info
    @oizone we have the same problem: we are an open source informatic school in limbe (cameroon), we are using and testing docker and docker-compose for some apllications, but we could not install kubernetes succesfully. we follow the but it fails by kubeadm init --pod-network-cidr any idea?
    Ross Beazley
    @michelcameroon use k3s instead, it's full kubernetes but built for raspberry pi as a first class citizen
    Ross Beazley
    @oizone I recommend k3s over k8s
    thanks i have install k3s, still working in it
    Marcel Stör
    Not sure if it's a feature or a bug and if so with which component:
    Stefano Passador
    Hi everyone!
    I just installed HypriotOS on my Raspberry PI.
    Unfortunately, I am unable to connect with the default password, while I can ping the raspberry.
    Is it a known issue?
    Hi folks. Just installed the latest image on my Pi. It is running and I can access it. I thought that it would include a ton of docker images/containers. It does not seem to have anything other than raspian running docker. Am I missing something?
    Hi everybody, I wonder if there is some French speaker here interested by talking about HypriotOS in a French podcast? If so, don't hesitate to send me a DM on Twitter @Saphoooo
    Andreas Zeidler
    Hi! So first of all, I've installed HypriotOS yesterday and was impressed how easy it was, thank you for all that work! :) Now I seem to be stuck with a network issue with docker stack, though. I can run docker run -d -p 4000:80 hypriot/rpi-busybox-httpd on the Pi and then successfully use curl http://kurt.local:4000/ from another machine. However, if I try to deploy the same image in a stack using and docker stack deploy -c hello.yaml hello I'm getting "no suitable node (unsupported platform on 1 node)". And even more strange is a "pihole" deployment that does work using docker stack, i.e. I can curl from the Pi, but get a timeout when trying the same from another machine. Does any of you have an idea what I might be doing wrong? :)
    Andreas Zeidler
    Ah, okay. Now that I wrote it down I found the solution for one of the issues, of course! :) So when I run docker stack deploy --resolve-image=never -c hello.yaml hello I can use curl from the Pi, but not from the "outside"…
    Andreas Zeidler
    Hmm, when comparing the iptables for run and stack it seems there's a masquerading rule missing. And iptables --table nat -A POSTROUTING -d -p tcp -m tcp --dport 4000 -j MASQUERADE indeed fixes the issue. I'm not sure that's the right thing to do, though.
    Andreas Zeidler
    Ok, the problem was the default swarm address pool was conflicting with my local network… %)
    Running docker swarm init --default-addr-pool --default-addr-pool-mask-length 24 fixed things… have a good night!
    Has anyone gotten LUKS disk encryption to work on the hypriot image? I keep getting "cryptsetup: cryptsetup failed, bad password or options?".
    I am setting up kubernetes on 2 rpi's running hypriot. I can make the cluster and I see both nodes. Does anyone have any experience with using ROS in this env? I have tried the various setups found on the web but nothing seems to work. My main issue is getting roscore ( in a docker container) talking to another ROS talker/listener containers. Any help is approeciated
    endre domiczi
    I have a couple of RPi 2's and would like to use Docker on them.
    Can I use Hypriot ? I see that was "promised" a (long) while ago.
    Yes, that should be possible
    What's the best edge embedded container framework ? For a a edge application running on multiple units
    Like but I prefer to use open source tools
    Ernesto Navarro
    Hello people, I don't find anywhere how to upgrade from an old version of hypriot to the latest. Can someone help me out, please?
    after installing master i try to install worker with sudo k3s agent --server https://myserver:6443 --token ${NODE_TOKEN}
    they not appear with get node i see that the token file has server and not node like i see in some docu. if i check the worker with sudo k3s kubectl get node they appear as master, some idea?
    with this command i could manage one worker
    curl -sfL | K3S_URL= K3S_TOKEN=K1056a31fc457716b4b41f372ec20c183584a4bfcea69db7ed88903b3355125351b::server:0581527deec67c7939e3d8ef501f4b96 sh -
    but when get node i recieve master but the "worker" has no role, is ti normal?
    Ross Beazley
    @michelcameroon the workers don't need a role, sounds like you have a working setup
    To see all the pods currently running across the cluster type something like
    kubectl get pods --allnamespaces -o wide
    It might be --all-namespaces
    thanks get pods i get i think all helm metrics local svclb coredns traefik svclb on knode now i try to make a second node (worker) with the command sudo k3s agent --server https://myserver:6443 --token ${NODE_TOKEN} it dos not work now i try with curl -sfL | K3S_URL= K3S_TOKEN=K1056a31fc457716b4b41f372ec20c183584a4bfcea69db7ed88903b3355125351b::server:0581527deec67c7939e3d8ef501f4b96 sh -
    i will tell if it is not working
    ok i have now 1 master and 2 nodes
    hi @wilsonianb thanks for your reply, i have the same result but i wonder if you try the command of the k3s startpage sudo k3s agent --server https://myserver:6443 --token ${NODE_TOKEN} for me it was not working
    Michael Cooper

    I am getting an error after doing docker swarm init on my hypriot system v1.12.x the following is what I am seeing:

    $ docker node ls
    Error response from daemon: This node is not a swarm manager. Use "docker swarm init" or "docker swarm join" to connect this node to swarm and try again.
    HypriotOS/armv7: pirate@cfDOC6 in ~
    $ docker swarm init
    Error response from daemon: This node is already part of a swarm. Use "docker swarm leave" to leave this swarm and join another one.

    Then I leave the swarm recreat it and the same thing all over again. Anyone have any ideas?


    I can't manage do the user-data file working on cloud-init.

    I'm using a Raspberry Pi 3B+, and flashing the OS with Etcher on Win10. When I do the flash without any change on the boot, it work as needed via eth0, ssh working as well. I managed to use the wifi connection manualy too.

    When I just change the username setting in the user-data file, it works too, with the new username.

    But when I unquote some line (for exemple : # Install any additional apt packages you need here

    • ntp)
      Or just the wifi setting part, it doesn't work.

    1) the wifi doesn't manage to connect my router at all
    2) when I try ssh, the default password (unchanged) doesn't work. (access denied)

    I walk around for quite a bit all around the github issues and the internet but it is very frustating so far.

    Hope an rescue here.

    Thanks a lot.

    I (finaly) managed to solve my issue after a long read on the cloud-init doc.

    It may help for others :

    When I changed the user-data config file, I just deleted the # and it stayed a space before the cloud-init modules name.

    Correcting the YAML syntax solved obviously the problem but it wasn't clear at my eye at first.
    I thought I had just to uncomment the lines I need to use.

    Thank you for the work done.

    Tarlan Isaev (organic)
    Hi folks, I've already updated to python3.8 on my Jetson Nano, but now docker-compose seems broken and I couldn't reinstall it successfully like yesterday. Is there a way to overcome this issue? :(
    i have installed a master and 2 agents on raspberry pi3. i have install helm and install some stable/... like mysql, dokuwiki, but nothing is working. can somebody tell me an example which really work?
    Vijay Panchal
    How to install raspi-config in hypriot OS 64 bit ?
    Hello, I Have a question. During boot I have custom user-data file and edit with package_update and package_upgrade to true value. but I think the RPI doessnt execute this. I We must define a special configuration so that the rpi is connected to the web during boot ?
    but i put a static address on all machine
    master and agents,
    Arjen Verstoep
    Hi. I'm building a custom SD image (with image-builder-rpi) but my /boot partition is empty after the process. Any idea what could cause it?
    Arjen Verstoep
    And is there a way to speed up subsequent builds? Every attempt is taking so much time :P
    Arjen Verstoep
    Alright, running the process again fixed the /boot issue (I guess there was an error during the build script somewhere. But is there a way to speed things up?: )
    Jon Davis
    newbie here .. hypriot blog says a new version was just announced .. download link sent along a v1.11.0 version file. However, a week or so ago I had downloaded a v1.12.0 file, which, extracted, is dated January 7. The version number is higher but it is older? I'm confused.
    Jakob Kukla
    Hi, I wanted to use OpenMax for HW acceleration on my pi 4. Now I noticed that the devices (/dev/vcsm and /dev/vchiq) that normally come on a standard install with raspbian or even arch arm are missing. Has anyone had a similar issue? How can I get them if they are not there by default? thanks :)