-
Notifications
You must be signed in to change notification settings - Fork 608
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
⚠️ QEMU (homebrew) was broken on Intel: [hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
#1742
Comments
[hostagent] Driver stopped due to error: "exit status 255"
[hostagent] Driver stopped due to error: "exit status 255"
Lima project probably needs to have an optional QEMU installation tarball (in a different project, like alpine-lima). |
[hostagent] Driver stopped due to error: "exit status 255"
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
)
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
)[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
The issue is now solved: |
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
Still seeing this when I run
Not sure if this is a lima or a colima problem. The VM does seem to start fine in spite of the warning. |
|
So it looks like it is a colima problem. Feel free to ignore! |
The binary seems broken again with the 8.1.0 bottle 😞 |
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
Besides the bloat, this seems like another good reason to not use Homebrew for the QEMU installation. |
I guess we should also rush to make (But when Homebrew fails to sign qemu binaries it may fail to sign lima binaries too, so switching to vz may not improve the situation) |
Adding a check for the entitlement to the brew formula would probably also be a good idea, to stop re-re-regressions from happening again in the future |
Already done in Lima v0.17.2: (The |
The Flux CI step often fails because the `attrs` module cannot be found on the system. This commit fixes this behaviour by explicitly installing the module through `pip` in the CI workflow. As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit adds an additional step to uninstall QEMU and recompile it from binary, which fixes the issue.
The Flux CI step often fails because the `attrs` module cannot be found on the system. This commit fixes this behaviour by explicitly installing the module through `pip` in the CI workflow. As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit adds an additional step to downgrade QEMU to v8.0.3, which fixes the issue.
The Flux CI step often fails because the `attrs` module cannot be found on the system. This commit fixes this behaviour by explicitly installing the module through `pip` in the CI workflow. As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit adds an additional step to downgrade QEMU to v8.0.3, which fixes the issue.
The Flux CI step often fails because the `attrs` module cannot be found on the system. This commit fixes this behaviour by explicitly installing the module through `pip` in the CI workflow. As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit adds an additional step to compile QEMU from source, which fixes the issue.
The Flux CI step often fails because the `attrs` module cannot be found on the system. This commit fixes this behaviour by explicitly installing the module through `pip` in the CI workflow. As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit adds an additional step to compile QEMU from source, which fixes the issue.
The Flux CI step often fails because the `attrs` module cannot be found on the system. This commit fixes this behaviour by explicitly installing the module through `pip` in the CI workflow. As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit upgrades the `douglascamata/setup-docker-macos-action` to version `v1-alpha.9` to overcome this issue.
The Flux CI step often fails because the `attrs` module cannot be found on the system. This commit fixes this behaviour by explicitly installing the module through `pip` in the CI workflow. As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit upgrades the `douglascamata/setup-docker-macos-action` to version `v1-alpha.9` to overcome this issue.
As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit upgrades the `douglascamata/setup-docker-macos-action` to version `v1-alpha.9` to overcome this issue.
As per issue lima-vm/lima#1742, the QEMU 8.1.0 Homebrew package is broken on Intel architectures. This commit upgrades the `douglascamata/setup-docker-macos-action` to version `v1-alpha.9` to overcome this issue.
The issue is solved again in Run |
It doesn't seem to be fixed for me.
Since I have this issue I can also no longer build containers. Here below the complete log
|
@janvda Using Lima without Colima may work as a workaround: Removing |
I was able to work around the issue by manually signing the wrapper link: Along with these binaries (which are signed, now, after I trod a path of destruction through all plausible binaries and therefore cannot un-sign to test and see whether signing them is needed): From my slightly-cleaned-up-for-display error below, something seems to be checking whether the wrapper itself is signed. On my system, the wrapper is a link to INFO[0000] starting ... context=vm
> Using the existing instance "colima"
> "QEMU binary
/Users/eq/.colima/_wrapper/4e1b408f843d1c63afbbdcf80c40e4c88d33509f/bin/qemu-system-x86_64
is not properly signed with the com.apple.security.hypervisor entitlement"
error="failed to run [
codesign --verify /Users/eq/.colima/_wrapper/4e1b408f843d1c63afbbdcf80c40e4c88d33509f/bin/qemu-system-x86_64
]: exit status 1 (out=
/Users/eq/.colima/_wrapper/4e1b408f843d1c63afbbdcf80c40e4c88d33509f/bin/qemu-system-x86_64:
code object is not signed at all
In architecture: x86_64
)"
> You have to sign the QEMU binary with the "com.apple.security.hypervisor" entitlement manually.
See https://github.com/lima-vm/lima/issues/1742 .
> [hostagent] Starting QEMU (hint: to watch the boot progress, see "/Users/ewehrwein/.lima/colima/serial*.log")
> SSH Local Port: 51848
> [hostagent] Waiting for the essential requirement 1 of 5: "ssh"
^C |
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
[hostagent] Driver stopped due to error: "signal: abort trap"
(or "exit status 255"
) ... QEMU has already exited
Hi, This issue should be reopened |
You are hitting something different (maybe https://gitlab.com/qemu-project/qemu/-/issues/1864 , which can be worked around by |
FYI The following workaround worked for me on my macbook pro with intel processors: |
@edw-eqix Beautiful - this is the current workaround for M1 macbooks until the maintainers release a fix |
I've just tried it now on my M1 mac, and this is the current output with the most recent
|
The story is different for Colima: The warning is negligible if the VM is working. The warning shouldn’t be printed if you use Lima directly without colima ( |
Thanks for the response, my bad, was on the wrong repo :) |
After signing the actual binary with the following command: codesign --sign - --entitlements entitlements.xml --force /usr/local/bin/qemu-system-$(uname -m | sed -e s/arm64/aarch64/) I also signed the wrapper with something like: codesign --sign - --entitlements entitlements.xml --force ~/.colima/_wrapper/4e1b408f843d1c63afbbdcf80c40e4c88d33509f/bin/qemu-system-x86_64 and it worked. Now It shows me message like this: INFO[0010] starting ... context=vm
> Using the existing instance "colima"
> QEMU binary "/Users/mutant/.colima/_wrapper/4e1b408f843d1c63afbbdcf80c40e4c88d33509f/bin/qemu-system-x86_64" seems properly signed with the "com.apple.security.hypervisor" entitlement
> [hostagent] Starting QEMU (hint: to watch the boot progress, see "/Users/mutant/.lima/colima/serial*.log")
> SSH Local Port: 60997
> [hostagent] Waiting for the essential requirement 1 of 5: "ssh"
|
As above, This hint (Option 3, local signing), worked for me. Intel Mac, Sonoma 14.1. |
Update (Aug 27, 2023): The issue is solved again in
8.1.0_1
(Homebrew/brew#15903 ,Homebrew/homebrew-core#140596Homebrew/homebrew-core#140643). Runbrew reinstall -f --force-bottle qemu
to install the updated v8.1.0 bottle.Update (Aug 23, 2023): This seems to be broken again in v8.1.0 😞 (Homebrew/homebrew-core#140244) . See the
Workarounds
below.Update (Aug 14, 2023): The issue is now solved in Homebrew/homebrew-core#139492 .
Run
brew reinstall -f --force-bottle qemu
to install the updated v8.0.4 bottle.Homebrew bottle of QEMU v8.0.4 (Intel)
iswas broken due to a signing issue: Homebrew/homebrew-core#139409(The error message can be also
[hostagent] Driver stopped due to error: "exit status 255"
)Workarounds
Option 1: Downgrade QEMU to v8.0.3
Option 2: Install QEMU from the source
Option 3: Sign the QEMU binary locally
Lima v0.17.2 shows a prompt to suggest applying this workaround.
Thanks to @z0sen for reporting this in abiosoft/colima#777
The text was updated successfully, but these errors were encountered: