Support for ARM based Docker images.

An ARM-based executor for ARM-based  Docker images would allow Raspberry Pi development as well as other ARM software.

  • Guest
  • Feb 22 2019
  • Taking votes
  • Attach files
  • Martin Jakubec commented
    04 Aug 07:00

    +1

  • Watt Iamsuri commented
    14 Jul 15:12

    +1

  • Henry Miskin commented
    16 Jun 13:53

    +100

  • Siying Dong commented
    08 Jun 22:08

    +1

  • Liz Fong-Jones commented
    15 May 19:10

    Amazon now has m6g, so this is only going to increase in importance...

  • Dan Pipe-Mazo commented
    09 May 01:59

    We built a CircleCI orb that supports using buildx as the builder for aarch64 and `amd64` : https://circleci.com/orbs/registry/orb/elementaryrobotics/atom?version=0.1.4

    Buildx is enabled by default in the CircleCI ubuntu 16.04 machine as of the 20200401 variant.

    See repo for release notes and usage examples:
    https://github.com/elementary-robotics/atom/tree/master/.circleci

    CircleCI team it would be much nicer if we had ARM64 servers to build on as building through QEMU is slow and painful.

  • Sri Ram commented
    27 Apr 01:19

    +1, This is blocking so many open source projects from support raspberry pi or ARM based devices. Please support ARM devices.

  • Martijn The commented
    24 Apr 13:13

    +1

  • Ian Delahorne commented
    21 Apr 20:41

    Hi,

    We'd like the ability to build and test on arm64, as we'd like to use the upcoming AWS m6g instance types to lower our costs.

  • Martin Grigorov commented
    26 Mar 08:55

    I was able to create a workflow based on machine executors and QEMU to build deb/rpm/apk packages for Varnish Cache for amd64 and aarch64 architectures. It could be used for any architecture supported by QEMU. The PR could be seen at https://github.com/varnishcache/varnish-cache/pull/3263

  • Emilio Fernandes commented
    18 Mar 13:28

    Many software shops are testing ARM64 servers these days. Latest ARM64 servers have similar performance as amd64 ones but consume less power, i.e. smaller bills at the end of the month.

    It would be awesome if CircleCI adds support for ARM64 architecture!

    TravisCI supports them for some time now. And Github Actions also recently added support for remote agents on ARM64 - https://github.blog/changelog/2019-12-03-github-actions-self-hosted-runners-on-arm-architectures/

    I am tasked to release arm64 packages for our software and now I will have to duplicate our config.yml into .travisci.yml for this reason.

  • Martin Grigorov commented
    18 Mar 09:14

    It would be cool if the following was possible: https://gist.github.com/martin-g/fc8ecab291796439157015e2f5c24843, i.e. a machine job to setup QEMU, and then any following Docker job to be executed in the machine VM

  • Martin Grigorov commented
    17 Mar 12:46

    +1 for supporting ARM64 architecture

  • Benjamin Zaporzan commented
    13 Mar 16:21

    +1

  • David Simmons commented
    21 Jan 23:20

    ARM support is currently a blocker for us to officially support ARM platforms. 

  • Takahiro Tanaka commented
    16 Jan 18:08

    AWS provides Arm EC2 instances and their cost performance is very good.

    We are considering to use them on our systems but CircleCI is a blocker for that because we use CircleCI for not only just testing but as a part of our deployment process for example build a package.

  • T L commented
    14 Jan 11:23

    ^This.  125 Billion ARM devices in the world - the software for which needs ARM support if it is developed under CI. Emulation via qemu is too slow.

  • Guest commented
    December 31, 2019 22:32

    +1

  • Varun Avashia commented
    November 28, 2019 16:05

    +1 : Yes, please

  • Pablo Molina commented
    October 31, 2019 16:57

    +1 
    I'm using Raspbian, my use case is running the tests, build the dist folder and pushing everything on a different repository.
    The goal is to have a ready to use application once checked out on my raspi.
    As I required the same arch to be run directly I would love to see this feature.

  • Load older comments