| JanC | I think shan3 wants to boot slower ;) | 00:40 |
|---|---|---|
| Caesar | Is there a way to validate an upstart job for anatomical correctness but otherwise not do anything with it? | 01:48 |
| shan3 | In Ubuntu 10.04, is it possible that a bash script I run from an Upstart job can exit with errors because it is not compliant with dash? | 03:28 |
| Keybuk | your question doesn't make sense | 03:30 |
| Keybuk | scripts run from an Upstart job (within the script...end script block) are run with /bin/sh, which on Ubuntu is dash | 03:32 |
| Keybuk | this means you *cannot* include so-called bashisms in them | 03:32 |
| Keybuk | they are not bash scripts | 03:32 |
| Keybuk | if you instead call out to an executable bash script using exec, then it will be run with whatever the #! line is at the top | 03:32 |
| Keybuk | (ie. set that to #!/bin/bash *not* #!/bin/sh) | 03:33 |
| shan3 | Keybuk: Thanks for the reply... but I cannot figure out why my bash script will not run from the Upstart job but works perfectly when run from terminal | 03:34 |
| Keybuk | can you pastebin the upstart job? | 03:34 |
| shan3 | sure... 1 sec | 03:34 |
| shan3 | http://pastebin.com/JVLiFwaZ | 03:36 |
| Keybuk | /usr/sbin/mfmrootdaemon being the bash script? | 03:36 |
| shan3 | yes | 03:36 |
| Keybuk | well, it could be any number of issues | 03:36 |
| Keybuk | for example, is /usr even mounted/ | 03:37 |
| Keybuk | it's stopped when rcS is started, which might be quite soon after udev is started | 03:37 |
| Keybuk | it might need mounted filesystems like /proc, which aren't available when udev is started | 03:37 |
| Keybuk | have you tried adding set +x to your bash script, and "console output" to the job, to see what it fails on? | 03:37 |
| shan3 | but it should work if I run 'sudo start myjob' , right? | 03:38 |
| Keybuk | if it works when you do that, but doesn't work otherwise, then it would be a dep issue | 03:38 |
| Keybuk | if it doesn't work when you "sudo start", then I'd do the set +x/console output trick to see how it fails | 03:38 |
| shan3 | no it doesn't work when I use start... I'll try set +x now | 03:39 |
| Keybuk | the set =x goes at the top of mfmrootdaemon obviously | 03:39 |
| Keybuk | set +x I mean | 03:39 |
| Keybuk | it makes bash print each line before executing it | 03:39 |
| Keybuk | so you'll see where it breaks | 03:39 |
| shan3 | Keybuk: do you mean set -x ? coz that print the commands and arguments... either way, I've added 'console output' ans set +x and also set -x... but still no output | 04:05 |
| Keybuk | err yes | 04:05 |
| Keybuk | maybe -x | 04:05 |
| shan3 | -x enables and +x disables... | 04:07 |
| shan3 | start outputs 'myjob start/running' but there is no sign on it in ps | 04:08 |
| shan3 | *of it | 04:08 |
| === Ng_ is now known as Ng | ||
| === ion_ is now known as ion | ||
| === h\h is now known as haraldh | ||
| Keybuk | notting: around? | 16:47 |
| notting | sure, why not | 16:47 |
| Keybuk | I hear RedHat have written their own init replacement, and will be using that instead of Upstart | 16:48 |
| Keybuk | care to comment? | 16:48 |
| notting | (give me a few minutes) | 16:52 |
| Keybuk | :-) | 16:52 |
| notting | someone who is a RH employee is writing something in their spare time. i have not looked at what they've written so far | 16:56 |
| notting | there are no current plans to either ship it, or not ship it | 16:57 |
| Keybuk | they claim RH is paying them to write it | 16:59 |
| Keybuk | and that it's a foregone shoe into Fedora | 16:59 |
| notting | i'm not exactly comfortable attempting to talk about a third party in their absence. i'd prefer to let them speak for themselves. | 17:09 |
| notting | but nothing has been built for fedora, or proposed for fedora, or discussed in fedora in any way | 17:09 |
| mclasen | I'm sure it will be, in time | 17:10 |
| Keybuk | notting: ok ;) | 17:16 |
| Keybuk | so you don't know much about it | 17:16 |
| notting | i know of it/the basics about it | 17:19 |
| ion | iNIHt | 17:28 |
| ion | Nothing wrong with NIH if you have a good reason for it. What’s the reason for the RH init? | 17:29 |
| notting | no linux subsystem is complete unless there's a implementation that starts with 'g' and an implementation that starts with 'k'. init has neither. | 17:34 |
| Keybuk | it's all about the 'u' now | 17:46 |
| * notting waits for canonical to come out with their ibis mascot, then | 17:55 | |
| sadmac | Keybuk: I heard this rumor too. My reaction was the same as yours. | 18:40 |
| === aaron01 is now known as aaron01_lunch | ||
| Keybuk | sadmac: I had a reaction? :p | 21:33 |
| sadmac | Keybuk: yes. I asked notting :) | 21:34 |
| Keybuk | when I first heard about it (six months ago?) I didn't expect it to actually get finished, so yeah, I guess I'm surprised about that ;p | 21:34 |
| Keybuk | ah, that it's basically launchd? | 21:34 |
| sadmac | Keybuk: I know nothing about it | 21:34 |
| sadmac | Keybuk: where'd you hear it from? | 21:35 |
| Keybuk | various people leaking things to me, until finally Lennart came clean and sent me the draft announcement today | 21:40 |
| sadmac | Keybuk: its lennart then? | 21:40 |
| Keybuk | who else? :p | 21:40 |
| sadmac | Keybuk: I actually did think that earlier. | 21:40 |
| * sadmac wonders if he hasn't done enough damage | 21:41 | |
| Keybuk | damage? he did a great job on Avahi, finally killing that evil mdns responder crap from Apple | 21:42 |
| Keybuk | and PulseAudio killed esd | 21:42 |
| sadmac | Keybuk: Pulseaudio killed puppies too | 21:42 |
| Keybuk | I never said that ;) you said that ;) | 21:42 |
| sadmac | Keybuk: avahi... no opinion, I'll give you that one | 21:42 |
| sadmac | Keybuk: Whenever Ubuntu gets that magical patch that makes Pulse not suck upstream then I'll be ok (not your fault in this case. He won't take the damn thing.) | 21:43 |
| Keybuk | is that the "apt-get remove pulseaudio" patch? | 21:45 |
| sadmac | Keybuk: also, do you look at it as killing esd or cementing alsa? | 21:45 |
| JanC | pulseaudio works fine | 22:06 |
| JanC | except for 1 usage scenario maybe (a system daemon like mpd needing to play sound, but even that's solvable) | 22:10 |
| Keybuk | yeah, it's always worked ok for me | 22:11 |
| sadmac | right now this box will periodically produce static instead of sound unless I kill pulse. | 22:13 |
| sadmac | and that's for straight alsa apps | 22:13 |
| sadmac | if it goes through gstreamer it hardlocks | 22:13 |
| sadmac | busy-looping inside libpulse | 22:13 |
| JanC | your gstreamer uses pulse I hope? | 22:13 |
| sadmac | JanC: yes. It seems to use it in such a way that it hits the pulseaudio glitch harder, or hits a different glitch. | 22:14 |
| JanC | anyway, never seen that happen recently on Ubuntu (for at least 1 year) | 22:14 |
| JanC | may depend on the hardware though | 22:15 |
| sadmac | intel | 22:15 |
| sadmac | this box is on its second motherboard too. same problems as the last chipset on this one. | 22:15 |
| JanC | I think you mean HDA, intel doesn't make audio chips (anymore?) | 22:16 |
| sadmac | yeah. that's what I mean | 22:16 |
| === rberger_ is now known as rberger | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!