I'm afraid that this is very similar to #433 but I can see it on amd64, more specifically, on Debian's CI system called debci. Here, the tests are running in a qemu VM with 4G of ram and 2 cores.

It seems to be quite reliably to break, but all failures are fairly inconsistent. Example runs I've observed so far:

• https://ci.debian.net/packages/n/netavark/unstable/amd64/51297650/
• https://ci.debian.net/packages/n/netavark/unstable/amd64/51298757/
• https://ci.debian.net/packages/n/netavark/unstable/amd64/51298772/
• https://ci.debian.net/packages/n/netavark/testing/amd64/51297649/
• https://ci.debian.net/packages/n/netavark/testing/amd64/51310986/

I've managed to trigger it once on my laptop, by running the tests on four qemu machines in parallel. I've kept going running 5 machines in parallel, but could not reproduce anymore.

Any ideas/suggestions? Maybe I should increase the sleep statements in helpers.bash?