[systemd-devel] [HEADSUP] systemd-222 around the corner

[Ansgar Burchardt]

Hi,

Colin Guthrie <gmane at colin.guthr.ie> writes:
> Bit late now, but does anyone else have the issue that calling
> "systemctl reload" will kill off any machines registered in machined.
>
> To me the problem commit is 9b420b3cfb8b93daf50e4cdbc92b05f2209ef893.
>
> machined: make sure to track machine unit states properly
>
> Can someone confirm this? (mentioned it in the "getent hosts <machine>"
> thread but no one confirmed - or denied - the problem there).

I can confirm that "systemctl daemon-reload" will make machined lose
track of containers (though the container itself keeps running) and
already opened an issue[1].

I haven't tried if reverting the commit you mentioned fixes it, but I
think the real issue might be systemd emitting UnitRemoved signals on
reload... It seems semantically incorrect to do so.

Ansgar

  [1] <https://github.com/systemd/systemd/issues/376>