Skip to content
Snippets Groups Projects
New look: Off

qemu: uncaught target signal 6 (Aborted) - core dumped on Apple Silicon M1 arm64

    • Open Issue created by Richardson Dx @richardsondx

    I'm getting the following error when I'm running docker-compose up and according to this docker thread this is a qemu related bug: https://github.com/docker/for-mac/issues/5148#issuecomment-784992663

    Here's the error:

     > [5/6] RUN bundle install --jobs 4 --path .bundle/:
#10 4.104 Fetching gem metadata from https://rubygems.org/./usr/local/lib/ruby/gems/2.3.0/gems/bundler-1.17.1/lib/bundler/worker.rb:90: [BUG] Segmentation fault at 0x00004318cbf428
#10 4.110 ruby 2.3.4p301 (2017-03-30 revision 58214) [x86_64-linux]
#10 4.110
#10 4.111 SEGV received in SEGV handler
#10 4.111 -- Control frame information -----------------------------------------------
#10 4.111 SEGV received in SEGV handler
#10 4.111 qemu: uncaught target signal 6 (Aborted) - core dumped
#10 4.125 Aborted

    I've passed the following config flag platform: linux/amd64 to my docker-compose.yml to try to resolve this with no luck.

    This issue is happening on the following environment

    • Macbook Pro M1 2020
    • Mac OS Big Sur 11.2.3
    • docker-compose version 1.29.1
    • docker engine: 20.10.6
    • docker desktop 3.3.3(64133)

    Do you know why this is happening and how could I fix it? and Is there an arm based build

    Edited
    To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information

    Child items ...

    • Activity

    • Richardson Dx changed the description

      changed the description

    • John Snow added hostos: macOS label

      added hostos: macOS label

    • John Snow added target: arm label

      added target: arm label

    • John Snow
      John Snow @jsnow
      Reporter

      @richardsondx Do you happen to be able to provide the following bits of information:

      • QEMU version
      • QEMU command line
      • A core dump from QEMU, if at all possible? (Or backtrace output that might suggest the failure)
      • Instructions to reproduce?
    • John Snow added workflowNeeds Info label

      added workflowNeeds Info label

    • Richardson Dx changed title from qemu: uncaught target signal 6 (Aborted) - core dumped on Apple Silicon M1 to qemu: uncaught target signal 6 (Aborted) - core dumped on Apple Silicon M1 arm64

      changed title from qemu: uncaught target signal 6 (Aborted) - core dumped on Apple Silicon M1 to qemu: uncaught target signal 6 (Aborted) - core dumped on Apple Silicon M1 arm64

    • Richardson Dx
      Richardson Dx @richardsondx
      Author

      @jsnow thank you for the prompt reply. I'm not sure because qemu isn't mentioned anywhere in my app. But this article: https://www.docker.com/blog/multi-arch-images/ mentions that it come packaged with Docker Desktop.

      If that's the case that could* be the version:

      % brew info qemu
qemu: stable 5.2.0 (bottled), HEAD

      Steps to reproduce:

      1. Install docker on an Apple M1 chip machine / arm64
      2. Set dockerfile.yml
      FROM { private url }
ADD Gemfile /app/Gemfile
ADD Gemfile.lock /app/Gemfile.lock
RUN bundle install
      1. Set docker-compose.yml
      version: "3"
services:
  web:
    platform: linux/amd64
    ...
      1. Run docker-compose up

      Updates

      • ( Thu, Jun 3 2021) – I was only getting that issue when using a a custom FROM which is running additional commands. One of the package in that repo might be causing the issue but its hard to pinpoint which one.
      Edited by Richardson Dx
    • Richardson Dx
      Richardson Dx @richardsondx
      Author

      @jsnow Backtrace output:

      #11 4.084 [BUG] Segmentation fault at 0x0000418d41e2a0
#11 4.085 ruby 2.3.4p301 (2017-03-30 revision 58214) [x86_64-linux]
#11 4.085
#11 4.085 -- Control frame information -----------------------------------------------
#11 4.086 c:0006 p:---- s:0017 e:000016 CFUNC  :pop
#11 4.087 c:0005 p:0010 s:0014 e:000013 BLOCK  /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:55 [FINISH]
#11 4.088 c:0004 p:---- s:0011 e:000010 CFUNC  :loop
#11 4.089 c:0003 p:0009 s:0008 e:000007 METHOD /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:54
#11 4.089 c:0002 p:0010 s:0004 e:000003 BLOCK  /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:88 [FINISH]
#11 4.090 c:0001 p:---- s:0002 e:000001 (none) [FINISH]
#11 4.090
#11 4.090 -- Ruby level backtrace information ----------------------------------------
#11 4.090 /usr/local/bin/bundle:22:in `<main>'
#11 4.090 /usr/local/bin/bundle:22:in `load'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/exe/bundle:37:in `<top (required)>'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/friendly_errors.rb:130:in `with_friendly_errors'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/exe/bundle:49:in `block in <top (required)>'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/cli.rb:24:in `start'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/vendor/thor/lib/thor/base.rb:485:in `start'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/cli.rb:30:in `dispatch'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/vendor/thor/lib/thor.rb:392:in `dispatch'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/vendor/thor/lib/thor/invocation.rb:127:in `invoke_command'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/vendor/thor/lib/thor/command.rb:27:in `run'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/cli.rb:258:in `install'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/settings.rb:134:in `temporary'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/cli.rb:259:in `block in install'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/cli/install.rb:64:in `run'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/installer.rb:24:in `install'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/installer.rb:72:in `run'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/process_lock.rb:9:in `lock'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/process_lock.rb:9:in `open'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/process_lock.rb:12:in `block in lock'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/installer.rb:84:in `block in run'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/installer.rb:254:in `ensure_specs_are_compatible!'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/definition.rb:183:in `specs'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/spec_set.rb:82:in `materialize'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/spec_set.rb:82:in `each'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/spec_set.rb:87:in `block in materialize'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/source/rubygems.rb:293:in `double_check_for'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/source/rubygems.rb:424:in `fetch_names'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/source/rubygems.rb:424:in `each'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/source/rubygems.rb:427:in `block in fetch_names'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher.rb:117:in `specs_with_retry'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/retry.rb:30:in `attempt'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/retry.rb:40:in `run'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher.rb:118:in `block in specs_with_retry'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher.rb:133:in `specs'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher.rb:133:in `each'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher.rb:134:in `block in specs'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher/compact_index.rb:16:in `block in compact_index_request'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher/compact_index.rb:16:in `call'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher/compact_index.rb:30:in `specs'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher/compact_index.rb:43:in `specs_for_names'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/compact_index_client.rb:70:in `dependencies'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher/compact_index.rb:99:in `block in parallel_compact_index_client'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher/compact_index.rb:99:in `each'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/fetcher/compact_index.rb:99:in `block (2 levels) in parallel_compact_index_client'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:36:in `enq'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:86:in `create_threads'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:86:in `new'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:86:in `initialize'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:88:in `block in create_threads'
#11 4.091 /usr/local/bundle/gems/bundler-2.2.17/lib/bundler/worker.rb:88:in `start'
#11 4.091
#11 4.091 -- Machine register context ------------------------------------------------
#11 4.092  RIP: 0x000000400215b1ca RBP: 0x0000000000000002 RSP: 0x000000401e398d50
#11 4.092  RAX: 0x000000002e382e2f RBX: 0x0000000000000020 RCX: 0x0000000000000024
#11 4.092  RDX: 0x000000401b807128 RDI: 0x000000401e399a20 RSI: 0x000000401e3996f8
#11 4.092   R8: 0x0000004003400000  R9: 0x000000401e398460 R10: 0x000000401e398410
#11 4.092  R11: 0x00000040034126c8 R12: 0x0000000000000020 R13: 0x000000400238a7a0
#11 4.092  R14: 0x000000401b806000 R15: 0x0000000000000020 EFL: 0x0000000000000202
#11 4.092
#11 4.092 -- C level backtrace information -------------------------------------------
#11 4.092 qemu: uncaught target signal 11 (Segmentation fault) - core dumped
#11 4.105 Segmentation fault
      Edited by Richardson Dx
    • John Snow
      John Snow @jsnow
      Reporter

      I don't have a setup where I can attempt to reproduce this myself (Don't have the right HW) and I am not familiar with how Docker for Desktop operates, but it would be very useful to see the stack trace from QEMU -- that's just the backtrace from the ruby component that started QEMU.

      Normally on linux, I'd say ulimit -c unlimited to allow for core dumps of any size to be created, and then coredumpctl debug to bring up a debugging interface for the most recent crash dump. No clue how this works on macOS.

      Maybe https://en.wikipedia.org/wiki/List_of_macOS_components#Crash_Reporter can help -- I am seeing a few StackOverflow links that suggest you need to use the same 'ulimit -c unlimited' invocation to allow core dump files of any size to be created in /cores, but from there I'm not 100% sure on how to get the stack trace.

      Usually: gdb ./path/to/qemu ./path/to/corefile followed by, in the GDB interface thread apply all bt. But you need to get the right executable there. It's probably x86_64-linux-user or i386-linux-user. QEMU ships a few different binaries for different purposes and Docker might actually even ship their own from what I'm reading. See if there are QEMU binaries located in the installation files for the Docker package(s). It should be fairly evident if you've picked the right one based on the output from GDB when you try to open it against the core file.

      This goes a little bit beyond initial triage for me, so I can't promise my own attention on this bug.

      Lastly, it'd be helpful if there's some way to coax Docker into printing or logging the QEMU CLI somehow. If it doesn't crash right away, or if there's a way to somehow delay the crash, you might be able to look at your process list while it runs to see the CLI args it was given, but that might not be feasible. Sorry, you might need to ask the Docker folks how to get this information to report it back to us.

    • John Snow added linux-user label

      added linux-user label

      • Marcel Blöcher
        Marcel Blöcher @dizlike

        Looks like I have a similar issue with Erlang.

        On a Mac M1 Air with Os Big sur 11.3.1

        Docker version 20.10.6, build 370c289
QEMU emulator version 5.2.0 (v5.2.0-11-ge7395802aa)

        Native arm images work as expected

        docker run -it --rm erlang:24 erl
docker run --platform linux/arm64/v8 -it --rm erlang:24 erl

        the following crashes

        docker run --platform linux/amd64 -it --rm erlang:24 erl
 # sys/unix/sys_signal_stack.c:269:sys_sigaltstack(): Internal error: Failed to set alternate signal stack
 # qemu: uncaught target signal 6 (Aborted) - core dumped

        I don't want to hijack this issue. Let me know if I can help out here - e.g. attaching a core dump.

      • John Snow
        John Snow @jsnow
        Reporter

        Too early to say if it's the same root cause. The environment is similar, but qemu-user could be crashing for a wide variety of different reasons. Different binaries being executed in the container may exercise different syscalls that we may fail to translate for different reasons. All I know so far is that "QEMU is crashing."

        I still don't have a Mac or an ARM machine at all, so -- I am just the tier 1 triage here -- I can't really explore this more in depth myself.

        The core dump by itself may not be useful without knowing precisely which version of the qemu binary actually crashed. See above for my response to Richardson Dx -- Can you try to identify exactly which QEMU binary is being launched by Docker and try to use GDB to obtain a stack trace from the core file? The Command line it crashed with is also a crucial piece of information.

      • Please register or sign in to reply
      • Richardson Dx
        Richardson Dx @richardsondx
        Author

        @jsnow I understand. I've shared some additional information below. And Thanks I'll try GDB! this is a great tool!

        I think this is the QEMU Binary being launched by docker:

        $ cd /Applications/Docker.app
$ sudo find . -type f -name "qemu"
./MacOS/qemu-system-aarch64

        I'm still not sure about:

        • QEMU Command Line
        • QEMU version
        • and getting the QEMU dump. <-- i'll check if the Docker Desktop app shows additional information.

        I found a way around this issue which I shared here – but it doesn't actually resolve our qemu issue.

        Edited by Richardson Dx
      • John Snow
        John Snow @jsnow
        Reporter

        Sorry to be helpless here -- The original docker issue you filed was closed and locked. Can you re-file a bug and ask them to help you obtain this information? You can point them here if you'd like; I'm willing to diagnose the problem but need some help to get the information we'd require. It will be helpful to know in the future as I am sure we will begin to see more and more M1 users finding their way to us.

      • Richardson Dx
        Richardson Dx @richardsondx
        Author

        @jsnow Ok sounds good, I will run docker and share the diagnostic error with them when I re-open the issue.

        @dizlike This might be related to your issue: I recently noticed that when I run my Elixir app under docker I was getting

        qemu: uncaught target signal 6 (Aborted) – core dumped

        Then when I tried to run the app on my machine locally I got a similar error:

        $ iex -S mix phx.server
[run] Segmentation  fault: 11

        Note: I was getting that error when I was installing erlang and elixir via brew with the latest versions. I then started using a version manager asdf and installed it on my M1 with Rosetta

        alias a="arch -x86_64"
a asdf install erlang 23.0
a asdf install elixir 1.8.1

        if you ran into issues while installing these versions; for example cannot find required auxiliary files: install-sh config.guess config.sub . you might have to reinstall autoconf. This is what I did to make it work:

        alias ibrew="arch -x86_64 brew"
ibrew install autoconf@2.69 && \
ibrew link --overwrite autoconf@2.69 && \
autoconf -V

        To fix the issue in my case I had to reinstall erlang and elixir. In my cases the versions of elixir I needed was elixir:1.8.1 and the version of erlang that worked with it was erlang 23.0. I had some issues running mix with versions from 20-22

        Edited by Richardson Dx
      • Please register or sign in to reply
    • Hu Jialun mentioned in issue #1763

      mentioned in issue #1763

    Please register or sign in to reply