On 02/06/2014 09:43 AM, Aaron Bull Schaefer wrote: > On Wed, Feb 5, 2014 at 12:05 PM, Fons Adriaensen <fons@xxxxxxxxxxxxxx> wrote: >> I had the same problem with remote poweroff some months ago. >> The solution I used was to create a delayed systemd target. > > > That's exactly what I ended up doing as well when building Arch > virtual machines for Vagrant [1]. > > [1] https://github.com/elasticdog/packer-arch/commit/6f747f94258b35f74c4fccb11c066f2933de5dd7 > Thanks for everyone's response. All good points, there isn't any need for an immediate reboot, and my point wasn't to complain about the behavior. As I try and make friends with systemd, I like to learn why things do what they do. The poweroff.timer solution is attractive and I like the (subshell) approach (fast, simple and readily available after kernel updates). I don't pretend to understand all the nitty-gritty that plays into difficulty ironing out ordering between systemd-user-sessions.service and network.target processing, but slowly the basics are sinking in. -- David C. Rankin, J.D.,P.E.
