mirror of
https://github.com/morgan9e/systemd
synced 2026-04-14 00:14:32 +09:00
6a05abb9b4
As registering the container creates a scope which might not be cleaned up completely before we run a next command in the same container, causing intermittent test fails: [ 63.424739] TEST-13-NSPAWN.sh[4231]: + systemd-nspawn --directory=/var/lib/machines/TEST-13-NSPAWN.sanity.zH2 bash -xec '[[ $USER == root ]]' [ 63.427504] systemd-nspawn[4381]: ░ Spawning container TEST-13-NSPAWN.sanity.zH2 on /var/lib/machines/TEST-13-NSPAWN.sanity.zH2. [ 63.437154] systemd[1]: Started TEST-13-NSPAWN.sanity.zH2.scope - Container TEST-13-NSPAWN.sanity.zH2. [ 63.437765] systemd-machined[1164]: New machine TEST-13-NSPAWN.sanity.zH2. [ 63.440311] TEST-13-NSPAWN.sh[4381]: + [[ root == root ]] [ 63.442046] systemd[1]: TEST-13-NSPAWN.sanity.zH2.scope: Killed unit cgroup '/machine.slice/TEST-13-NSPAWN.sanity.zH2.scope' with SIGKILL on client request. [ 63.442583] systemd-nspawn[4381]: Container TEST-13-NSPAWN.sanity.zH2 exited successfully. [ 63.443073] systemd-machined[1164]: Machine TEST-13-NSPAWN.sanity.zH2 terminated. [ 63.448728] TEST-13-NSPAWN.sh[4231]: + systemd-nspawn --directory=/var/lib/machines/TEST-13-NSPAWN.sanity.zH2 --user=testuser bash -xec '[[ $USER == testuser ]]' [ 63.451209] systemd-nspawn[4385]: ░ Spawning container TEST-13-NSPAWN.sanity.zH2 on /var/lib/machines/TEST-13-NSPAWN.sanity.zH2. [ 63.455295] systemd-nspawn[4385]: Failed to allocate scope: Unit TEST-13-NSPAWN.sanity.zH2.scope was already loaded or has a fragment file. [ 63.456139] systemd[1]: TEST-13-NSPAWN.sanity.zH2.scope: Deactivated successfully. [ 63.461292] TEST-13-NSPAWN.sh[2839]: + at_exit Since even systemd-nspawn's man page suggests not to register containers with systemd-machined if they don't run a service manager, let's do just that to mitigate the race. Resolves: #39629