NO TO SYSVINIT - or initscripts, rather!

Posted Nov 5, 2018 23:24 UTC (Mon) by luto (guest, #39314)
In reply to: NO TO SYSVINIT - or initscripts, rather! by Cyberax
Parent article: Init system support in Debian

But what if I actually *want* systemd to manage my services?

NO TO SYSVINIT - or initscripts, rather!

Posted Nov 5, 2018 23:28 UTC (Mon) by Cyberax (✭ supporter ✭, #52523) [Link] (4 responses)

It will be able to do it, just without any resource controllers. And you can have a separate cgroups tree with controllers attached.

NO TO SYSVINIT - or initscripts, rather!

Posted Nov 12, 2018 17:22 UTC (Mon) by nix (subscriber, #2304) [Link] (3 responses)

I thought that in these days of cgroupsv2 there was a single unified tree, which systemd insists on managing?

NO TO SYSVINIT - or initscripts, rather!

Posted Nov 12, 2018 19:54 UTC (Mon) by Cyberax (✭ supporter ✭, #52523) [Link] (2 responses)

You can still have multiple trees. The central tree with controllers attached and multiple v1-like trees without controllers. This is enough for systemd to manage the processes.

NO TO SYSVINIT - or initscripts, rather!

Posted Nov 13, 2018 0:18 UTC (Tue) by nix (subscriber, #2304) [Link] (1 responses)

Oh, so the only restriction with v2 is that *controllers* can only appear on one unified tree? You can have lots of other trees for the sole purpose of defining relationships between processes?

That's... useful. As in, I think I'll have a use for it in my own work in the next week. :) I wish I'd realised v2 worked this way years ago!

NO TO SYSVINIT - or initscripts, rather!

Posted Nov 13, 2018 0:21 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link]

Correct. You can still use a thin tree only for processing tracking.