Re: [systemd-devel] Targeting `systemd --user` for a distributed app without root access: dead-end or neat?

2016-04-08 Thread sbaugh
Hugues Malphettes writes: > Hi, > > We are discussing adding support for coreos's fleet to control a user > session of systemd. > On paper it is perfect to run a distributed application in an > infrastructure without root access. This would be extremely helpful for some projects of mine. For a w

Re: [systemd-devel] Monitor path to commit to git

2016-04-08 Thread Josh Triplett
Lennart Poettering wrote: > inotify doesn't really provide such a feature, and fanotify is > crap. Leaving aside any other issues with fanotify, it doesn't seem to provide this feature either; "man fanotify" says "Fanotify monitoring of directories is not recursive: to monitor subdirectories under

Re: [systemd-devel] udev removing and re-adding partition /dev entries after read-write open close of whole disk device

2016-04-08 Thread Mike Fleetwood
On 18 March 2016 at 17:43, Mike Fleetwood wrote: > Hi, > > I've noticed that since approximately udev version 219, opening a whole > disk device entry read-write, then closing it triggers udev to remove > and re-create all the partition /dev entries. > > Because of this GParted (graphical disk par

Re: [systemd-devel] udev vs. nscd vs. /var automount

2016-04-08 Thread Michal Soltys
On 2016-04-08 18:12, Andrei Borzenkov wrote: 08.04.2016 18:20, Michael Biebl пишет: 2016-04-08 16:25 GMT+02:00 Michal Soltys : On your root partition keep /var/run symlink to /run - so regardless if /var is or isn't mounted, the path will be correct. Wouldn't that help ? (unless I missed someth

Re: [systemd-devel] udev vs. nscd vs. /var automount

2016-04-08 Thread Michael Biebl
2016-04-08 18:12 GMT+02:00 Andrei Borzenkov : > 08.04.2016 18:20, Michael Biebl пишет: >> 2016-04-08 16:25 GMT+02:00 Michal Soltys : >>> On your root partition keep /var/run symlink to /run - so regardless if /var >>> is or isn't mounted, the path will be correct. Wouldn't that help ? (unless >>> I

Re: [systemd-devel] udev vs. nscd vs. /var automount

2016-04-08 Thread Andrei Borzenkov
08.04.2016 18:20, Michael Biebl пишет: > 2016-04-08 16:25 GMT+02:00 Michal Soltys : >> On your root partition keep /var/run symlink to /run - so regardless if /var >> is or isn't mounted, the path will be correct. Wouldn't that help ? (unless >> I missed something) > > That only work for late boot

Re: [systemd-devel] udev vs. nscd vs. /var automount

2016-04-08 Thread Michael Biebl
2016-04-08 16:25 GMT+02:00 Michal Soltys : > On your root partition keep /var/run symlink to /run - so regardless if /var > is or isn't mounted, the path will be correct. Wouldn't that help ? (unless > I missed something) That only work for late boot though. Consider the case where /var is a separ

Re: [systemd-devel] udev vs. nscd vs. /var automount

2016-04-08 Thread Mantas Mikulėnas
That won't work with autofs automounts – they overlay the root partition just like a regular mount would. So it's not that udev can't find the socket, but rather that it triggers autofs whenever it tries to connect. On your root partition keep /var/run symlink to /run - so regardless if > /var is

Re: [systemd-devel] udev vs. nscd vs. /var automount

2016-04-08 Thread Michal Soltys
On 2016-04-06 11:15, Łukasz Stelmach wrote: Hi, I've hit a problem caused by a mix of: automounting + glibc + udev + my partition layout. Apparently it is impossible to make /var automountable because udev (which needs to enumerate devices befor mounting them) is trying to connect to /var/run/ns

Re: [systemd-devel] Monitor path to commit to git

2016-04-08 Thread killermoehre
Am 07.04.2016 um 20:11 schrieb Florian Lindner: > Hello, > I want to have a unit that monitors a path and commits automatically to > git whenever something changes. It usually works, like that: > > # cat [email protected] > [Unit] > Description=Automatic commit for %f > > [Service] > Type = one

Re: [systemd-devel] Monitor path to commit to git

2016-04-08 Thread Lennart Poettering
On Thu, 07.04.16 20:11, Florian Lindner ([email protected]) wrote: > It basically works but has two issues: > > 1) The path unit does not seem to monitor the path recursively, > therefore I don't get a commit when a file in a subdirectory changes inotify doesn't really provide such a feature,