Improve cancel conversion failure message

[ben-grande]

I didn't notice that the failure message hardcoded the time last time I changed the timeout.

It's failing on openqa. Investigation says that it is using latest version fedora-43-xfce: qubes-pdf-converter-2.1.26-1.4.fc43.noarch. Might need to increase timeout to 30.

---

scheduled for openqa: QubesOS/qubes-core-admin#783 (comment)

[qubesos-bot]

OpenQA test summary

Complete test suite and dependencies: https://openqa.qubes-os.org/tests/overview?distri=qubesos&version=4.3&build=2026042816-devel&flavor=pull-requests

Test run included the following:

• Improve cancel conversion failure message #41 (b578c2c)
• DONOTMERGE - Enable qmemman debug log qubes-core-admin#783 (QubesOS/qubes-core-admin@d986e13)
• Improve policy admin client and related tools qubes-core-qrexec#223 (QubesOS/qubes-core-qrexec@67941b5)
• Use new exception for handling absent policy qubes-desktop-linux-manager#301 (QubesOS/qubes-desktop-linux-manager@57da930)

New failures, excluding unstable

Compared to: https://openqa.qubes-os.org/tests/overview?distri=qubesos&version=4.3&build=2026032404-devel&flavor=update

Failed tests

No failures!

Fixed failures

Compared to: https://openqa.qubes-os.org/tests/170766#dependencies

3 fixed

• system_tests_extra

  • system_tests: Fail (unknown)
    Tests qubes.tests.extra failed (exit code 1), details reported sepa...

  • system_tests: Failed (test died)
    # Test died: Some tests failed at qubesos/tests/system_tests.pm lin...

  • TC_01_InputProxyExclude_debian-13-xfce: test_000_qemu_tablet (error)
    qubes.exc.QubesVMError: Cannot connect to qrexec agent for 120 seco...

Unstable tests

Details

Performance Tests

Performance degradation:

No issues

Remaining performance tests:

No remaining performance tests

[marmarek]

It's failing on openqa. Investigation says that it is using latest version fedora-43-xfce: qubes-pdf-converter-2.1.26-1.4.fc43.noarch. Might need to increase timeout to 30.

I'm not sure if it's about about the timeout. If you watch the video, it looks like nothing is happening. And also, it worked on Debian. Might be something not working with the cancellation.
FYI @Jayant-kernel

[Jayant-kernel]

@marmarek
Thanks for awaring
Could the issue be that on Fedora 43, qrexec-client-vm exits from SIGINT before our cleanup runs, so terminate_proc skips it (returncode is already set) and the DispVM is left hanging?
Debian might just handle the timing differently.

[marmarek]

If qrexec-client-vm already exited, terminate_proc wouldn't do anything else anyway. Is there some other part of the cleanup I'm missing?

[Jayant-kernel]

Yes, I think you’re right. I was mixing up local cleanup with DispVM cleanup. Our code only cancels local tasks/temp files and terminates qrexec-client-vm if it is still running; the actual DispVM cleanup should happen on the @dispvm/qrexec side after the service ends.

So if qrexec-client-vm already exited but the DispVM remains, the issue is probably delayed/failed auto-cleanup, not terminate_proc.