| Varazir | start on <service has run before> | 10:20 |
|---|---|---|
| Varazir | can it be done ? | 10:20 |
| Varazir | nvm | 10:21 |
| SpamapS | Varazir: did you figure out what you needed? | 15:26 |
| Varazir | SpamapS: I think I solved the problem not exting my script | 15:27 |
| Varazir | I added and started htpc to the lightdm.conf file | 15:27 |
| Varazir | the orginal htpc.conf file had exit 0 | 15:27 |
| Varazir | and then lightdm wouldn't start due to the script wasn't running | 15:28 |
| SpamapS | ah ok | 15:30 |
| Varazir | SpamapS: Didn't work | 17:14 |
| Varazir | for some reason my script do not start | 17:15 |
| Varazir | init-checkconf /etc/init/htpc.conf | 17:17 |
| Varazir | ERROR: failed to ask Upstart to check conf file | 17:17 |
| Varazir | And I don't get any error msg what I can see | 17:25 |
| wakko | hi there | 21:19 |
| wakko | i am lookup for a default set of jobs that does not use rc, does someone knows where i could find that ? | 21:21 |
| wakko | slookup/looking/ | 21:21 |
| JanC | you mean without sysvinit compatibility? | 21:33 |
| wakko | yes | 21:33 |
| JanC | hm, I don't know about any default set like that | 21:33 |
| JanC | I think it's only used like that in embedded systems | 21:34 |
| wakko | actually i am trying to bake an embedded system | 21:35 |
| JanC | I think Palm/HP webOS uses it like that | 21:36 |
| wakko | thanks for the info | 21:36 |
| JanC | I'm not sure how Google ChromeOS uses it | 21:36 |
| wakko | for the moment i am stuck at the runlevel thing | 21:39 |
| JanC | wakko: maybe also ask on the mailing list | 21:39 |
| wakko | when i type runlevel it tells me "runlevel:/var/run/utmp: No such file or directory" | 21:39 |
| JanC | you don't have to use runlevels | 21:39 |
| wakko | i need to find out what does populate this file :) | 21:39 |
| wakko | i would like not to use them indeed, but at the same time i'd like to have the reboot command to work | 21:41 |
| JanC | "reboot --force" reboots the system without any need for runlevels | 21:45 |
| JanC | but you'll have to make sure everything is topped properly before doing that yourtself | 21:46 |
| JanC | *yourself | 21:46 |
| JanC | and *stopped* (sigh) | 21:46 |
| wakko | i don't understand why reboot is shipped with upstart and still relies on /etc/init.d/rc | 21:47 |
| wakko | (according to the cookbook) | 21:47 |
| JanC | there is also a reboot syscall BTW | 21:47 |
| JanC | wakko: because of the sysvinit compatibility I suppose | 21:47 |
| JanC | you don't want to reboot before your database that still uses a sysvinit script has written all data to disk, for example ☺ | 21:48 |
| wakko | i was thinking reboot would send an event to upstart thru dbus, and then upstart would call /etc/init.d/rc using a job | 21:49 |
| slangasek | that's correct | 21:50 |
| JanC | well, that's more or less what it does | 21:50 |
| slangasek | and then it's the init scripts that finish up tearing everything down, and calling the reboot syscall | 21:50 |
| JanC | if you don't use --force | 21:50 |
| slangasek | the reason for this is that event based systems make it much easier to respond to events quickly (on startup or shutdown or anywhere in between), but they make it *harder* to know when you're actually done processing | 21:51 |
| JanC | if you don't use the sysvinit compatibility, you'll have to do that syscall yourself (e.g. by calling "reboot --force") | 21:51 |
| wakko | JanC: i cannot use the reboot command without --force, so upstart still get a "reboot event" turns off all deamons and finally do a reboot --force ? | 21:52 |
| JanC | well, you can do the "reboot --force" (or similar another way) from an upstart job, of course | 21:53 |
| wakko | that would be nice | 21:53 |
| JanC | I'm sure distros that don't use sysvinit have their own implementation of reboot & shutdown BTW | 21:54 |
| JanC | e.g. Gentoo & CRUX | 21:54 |
| wakko | Does Gentoo use upstart ? | 21:54 |
| JanC | no, (certainly not by default) | 21:55 |
| JanC | but they don't use sysvinit either | 21:55 |
| JanC | they have their own init system | 21:55 |
| wakko | i think it's OpenRc by default | 21:55 |
| JanC | yes | 21:55 |
| JanC | which was originally derived from bsd init IIRC, but evolved a lot since | 21:56 |
| wakko | i think i could keep the runlevel stuff :) so i'll don't have to rewrite them and i can stay with upstart "builtins" | 21:56 |
| JanC | well, it would be interesting to see upstart working without sysvinit/runlevels/etc. | 21:57 |
| JanC | but I understand you probably have time constraints about getting this to work ;) | 21:58 |
| wakko | the cookbook says reboot does a "runlevel RUNLEVEL=6 PREVLEVEL=2" | 21:58 |
| wakko | i think i could keep the runlevel logic but skip the rc scripts | 21:58 |
| JanC | that would probably work | 21:59 |
| wakko | JanC, slangasek, thanks for your help | 21:59 |
| JanC | it's more or less what happens if /etc/init.d/* & /etc/rc.local are empty, I suppose ☺ | 22:00 |
| wakko | :) | 22:00 |
| wakko | maybe, but i still need to clarify some points so i can tell | 22:01 |
| wakko | lol it works now... | 22:02 |
| wakko | i'll definitely write an article or something later | 22:02 |
| wakko | gn all and thanks again | 22:06 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!