| [05:29] <pitti> Good morning | |
| === work_alkisg is now known as alkisg | |
| [06:32] <Unit193> pitti: I have a wily hardware computer that uses this same kernel with kdbus, so I could do some testing there too if it'd help. | |
| [06:40] <dholbach> good morning | |
| === marcusto_ is now known as marcustomlinson | |
| === vrruiz_ is now known as rvr | |
| === pete-woods1 is now known as pete-woods | |
| === greyback is now known as greyback|doc | |
| === alkisg is now known as work_alkisg | |
| === MacSlow is now known as MacSlow|lunch | |
| === greyback|doc is now known as greyback | |
| === MacSlow|lunch is now known as MacSlow | |
| === attente_ is now known as attente | |
| === pgraner is now known as pgraner-dr | |
| === _salem is now known as salem_ | |
| === zyga_ is now known as zyga | |
| === Malsasa_ is now known as Malsasa | |
| === henrix_ is now known as henrix | |
| [16:53] <hallyn> pitti: mbiebl_: hi, around? | |
| [16:53] <mbiebl_> I am | |
| [16:54] <hallyn> mbiebl_: cool! i wanted to ask if there is default behavior for systemd oneshot units where anything started by them is killed when the job is done? | |
| [16:54] <hallyn> in particular, | |
| [16:54] <hallyn> lxc-net.service is a type=oneshot which runs a script which starts a dnsmasq. | |
| [16:54] <hallyn> that dnsmasq seems to just be disappearing; but if i run the same command by hand as root it persists (as it should) | |
| [16:55] <mbiebl_> why is it oneshot if it starts a daemon? | |
| [16:55] <hallyn> well it sets up and configures a bridge | |
| [16:55] <hallyn> i didn't even consider making that a daemon, maybe it should... | |
| [16:55] <hallyn> https://github.com/hallyn/lxc/blob/testing/config/init/systemd/lxc-net.service.in is the current job fwiw | |
| [16:55] <mbiebl_> and how does it start the dnsmasq daemon? | |
| [16:56] <hallyn> and https://github.com/hallyn/lxc/blob/testing/config/init/common/lxc-net.in is the script | |
| [16:56] <mbiebl_> hallyn: if it starts a daemon, it shouldn't be a oneshot type service | |
| [16:56] <hallyn> The dnsmsq command is: Kdnsmasq -u lxc-dnsmasq --strict-order --bind-interfaces --pid-file=/run/lxc/dnsmasq.pid --conf-file=/dev/null --listen-address 10.0.3.1 --dhcp-range 10.0.3.2,10.0.3.254 --dhcp-lease-max=253 --dhcp-no-override --except-interface=lo --interface=lxcbr0 --dhcp-leasefile=/var/lib/misc/dnsmasq.lxcbr0.leases --dhcp-authoritative | |
| [16:56] <hallyn> mbiebl_: lemme try that. so it should be 'simple'? | |
| [16:57] <hallyn> or, 'forking'? /me reads | |
| [16:57] <mbiebl_> depends on what your daemon does | |
| [16:57] <hallyn> mbiebl_: so if i make it type=simple, and lxc starts After=lxc-net, | |
| [16:58] <hallyn> oh wait. i've made it worse than you think. | |
| [16:58] <hallyn> so the dameon is started during the ExecStartPre | |
| [16:58] <mbiebl_> type=simple has one downside | |
| [16:58] <hallyn> because it needs to complete before lxc, which is After=lxc-net, starts | |
| [16:58] <hallyn> uh, "completes" | |
| [16:58] <mbiebl_> the service will be considered started as soon as the process has been spawned | |
| [16:58] <mbiebl_> it doesn't wait for the service "to be up" | |
| [16:58] <hallyn> right | |
| [16:59] <hallyn> so what does starting it in ExecStartPre do with respect to that? | |
| [16:59] <mbiebl_> hallyn: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=778913 has some more details on that behaviour | |
| [16:59] <ubottu> Debian bug 778913 in openssh-server "openssh-server: init (at least systemd) doesn't notice when sshd fails to start and reports success" [Important,Open] | |
| [17:01] <mbiebl_> <hallyn> so what does starting it in ExecStartPre do with respect to that? → I don't quite understand that question | |
| [17:01] <hallyn> 1. wlil the daemon started during ExecStartPre get killed, | |
| [17:01] <hallyn> 2. will the service be deemed failed bc ther eis no ExecStart, | |
| [17:02] <hallyn> 3. will lxc then not start unitl the ExecStartPre script is complete | |
| [17:02] <mbiebl_> ExecStartPre is not supposed to be used to start long running daemon | |
| [17:02] <hallyn> maybe i'll just call the lxc-net start script from the 'lxc' execstartpre | |
| [17:03] <hallyn> that seems to make more sense | |
| [17:03] <mbiebl_> isn't that what you do today? | |
| [17:03] <mbiebl_> ExecStartPre=@LIBEXECDIR@/lxc/lxc-net start | |
| [17:03] <mbiebl_> that seems wrong to me | |
| [17:03] <mbiebl_> ExecStartPre is to to setup your daemon environment | |
| [17:04] <mbiebl_> the actual daemon should be started via ExecStart | |
| [17:04] <mbiebl_> from the systemd.service man page | |
| [17:04] <mbiebl_> Note that ExecStartPre= may not be used to start long-running processes. All processes forked off by processes | |
| [17:04] <mbiebl_> invoked via ExecStartPre= will be killed before the next service process is run. | |
| [17:06] <mbiebl_> hallyn: I suggest reading the systemd.service man page | |
| [17:06] <mbiebl_> hallyn: gotta run now | |
| [17:07] <hallyn> mbiebl_: thanks! | |
| [17:11] <hallyn> hm, that still causes the dnsmasq to get killed | |
| [17:11] <hallyn> apparently ExecStartPre cannot start long-running daemons | |
| [17:12] <hallyn> hrmp. "Note that ExecStartPre= may not be used to start long-running processes. All processes forked off by processes invoked via ExecStartPre= will be killed before the next service process is run." | |
| === pgraner-dr is now known as pgraner | |
| === pgraner is now known as pgraner-dr | |
| === salem_ is now known as _salem | |