As the summary says, systemd-nspawn fails to star a container when --tmpfs is used. If omitting only --tmpfs from the command line then the container starts successfully. I'm using the following options: /usr/bin/systemd-nspawn -M centos --link-journal=no --bind=/home/centos:/home --tmpfs=/tmp:mode=1777 --network-bridge=mybr -D /virt/containers/centos It fails with: Spawning container centos on /virt/containers/centos. Press ^] three times within 1s to kill container. Container centos failed with error code 1. OS: Arch Linux Systemd: 217-6
Hmm, something appears to fail in the container without a proper error message. Could you run the nspawn command within "strace -f" and see if that shows any particular errors in the client that might be causing this? If in doubt simply attach the strace -f fragment right before the child process in the container exits, and we can have a look. Thanks.
Created attachment 110412 [details] strace output Strace output is attached. Nothing seems to stand out as a possible source of the problem (to me at least), but I don't have much experience interpreting the output of strace (or debugging in general). Like I said before, if I omit the --tmpfs option I can start the container successfully. I can't say exactly when this stopped working as I don't use nspawn very often, but I haven't changed anything in my setup (systemd unit file or container) since the last time it worked fine.
Ah, the strace output was quite useful. Should be fixed in git now: http://cgit.freedesktop.org/systemd/systemd/commit/?id=04a919394069cf024559f78eb46692a3739641eb WOuld be great to give this test run, to verify that all is good now for you! Thanks!
It seems to be working as expected here. Thanks for fixing it :)
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.