| Caesar | # service autofs status | 00:18 |
|---|---|---|
| Caesar | autofs stop/killed, process 3690 | 00:18 |
| Caesar | # pgrep automount | 00:18 |
| Caesar | 3687 | 00:18 |
| Caesar | I think Upstart is getting very confused :-( | 00:19 |
| Caesar | # service autofs stop | 00:19 |
| Caesar | stop: Job has already been stopped: autofs | 00:19 |
| Caesar | http://paste.ubuntu.com/389279/ has my Upstart job definition, which I'm trying to test | 00:20 |
| Caesar | (on Ubuntu 10.04) | 00:26 |
| Keybuk | Caesar: bug #406397 | 00:53 |
| Caesar | Yarp | 00:56 |
| Keybuk | probably got confused following the pid chain and ended up with something that isn't actually autofs | 00:58 |
| Keybuk | like one of the fork parents | 00:58 |
| Caesar | Yeah | 01:07 |
| Caesar | "expect fork" after a clean boot fixed it | 01:07 |
| Keybuk | sweet | 01:12 |
| Keybuk | sorry about that one | 01:12 |
| Keybuk | I have a fix for it ;) | 01:12 |
| Keybuk | but it's a M one | 01:12 |
| Caesar | Doh | 01:12 |
| Caesar | It's pretty bad how busted things get in this situation | 01:12 |
| Caesar | start and stop just hang | 01:13 |
| Keybuk | yeah | 01:16 |
| Keybuk | it's because it's not supposed to be able to happen :p | 01:16 |
| Caesar | heh | 01:20 |
| Caesar | So what's currently doing my head in is the timing between the legacy init scripts and the upstart jobs | 01:21 |
| Caesar | Because a lot of the upstart jobs aren't runlevel dependent (maybe if I follow the dependency chain back I'll find that they are) | 01:22 |
| Keybuk | they're not supposed to be | 01:23 |
| Keybuk | there's a vague wishy-washy attempt to make runlevels go away :p | 01:23 |
| Caesar | heh | 01:23 |
| Keybuk | btw are you coming to the-woods-outside-Brussels? | 01:23 |
| Caesar | Sadly not, as my wife is due on May 24 | 01:24 |
| Caesar | And I'm transferring off Goobuntu after my paternity leave, so I won't have a justification for coming to future UDSes | 01:24 |
| Keybuk | aww :-( where are you transferring to? | 01:24 |
| Caesar | The security team | 01:24 |
| Keybuk | before you go, I'd really appreciate it if I could get a kind of "what Google would like from Upstart" mail | 01:25 |
| Keybuk | what you've found difficult about it | 01:25 |
| Keybuk | problems you've had | 01:25 |
| Keybuk | what you wish it did | 01:25 |
| Keybuk | what you wish it did better | 01:25 |
| Keybuk | etc. | 01:25 |
| Caesar | I can do that, particularly after we've deployed 10.04 | 01:25 |
| Caesar | Right now I'm just having a hell of a time debugging the boot process. That's due to a combination of factors though | 01:25 |
| Keybuk | yeah, as I think I've talked to you before, I want Upstart to keep an event log | 01:26 |
| Keybuk | so after a boot you can get a log like | 01:27 |
| Keybuk | "the foo event happened, so I started apache, but I couldn't start squid yet because it was still waiting on apache" | 01:27 |
| Keybuk | "the started apache happened, so I started squid" | 01:27 |
| Caesar | That sounds perfect | 01:27 |
| Keybuk | and want that to be hooked up in a way that on a hung system, you can have upstart squirt it out to netconsole | 01:27 |
| Keybuk | "dump my brain" | 01:27 |
| Caesar | The startup event is the root of everything isn't it? | 01:28 |
| Keybuk | yes | 01:30 |
| Keybuk | though in later versions you may be able to do "on system starting" and "on system started" | 01:30 |
| Keybuk | as two separate bits | 01:30 |
| Caesar | Okay, so startup -> mountall -> (local-filesystems) -> (run level 2) -> rc 2 | 01:31 |
| Caesar | Meanwhile, (local-filesystems) -> networking | 01:31 |
| Caesar | So it sounds like the rc 2 legacy init scripts are going to run in parallel with the network coming up | 01:31 |
| Caesar | Keybuk: something my cubemate was just lamenting is how restart isn't "restart or start" | 01:49 |
| Keybuk | Caesar: that's deliberate ;-) | 01:52 |
| Caesar | I thought as much | 01:52 |
| Caesar | Is there a "start or restart" equivalent? | 01:52 |
| Keybuk | I really hate how "/etc/init.d/foo restart" can start a service that wasn't running to begin with | 01:52 |
| Keybuk | stop foo ; start foo | 01:52 |
| Caesar | Right | 01:52 |
| Caesar | A bit more typing | 01:53 |
| sadmac | Keybuk: I thought there was a restart. moreover I thought it was slightly different than stop;start | 01:55 |
| sadmac | or was that for 1.0 | 01:55 |
| Keybuk | there is a restart | 01:55 |
| Keybuk | and it is slightly different ;) | 01:55 |
| Keybuk | Caesar had just asked whether there was an alternative to it that wasn't slightly different to stop;start ;) | 01:56 |
| sadmac | Keybuk: I'd think the preference would be restart foo || start foo | 02:17 |
| Keybuk | that'd work too ;) | 03:44 |
| sadmac | Keybuk: I have an ubuntu user who says he would like to "bitch slap ubuntu for making it look like a friggin mac" and would like me to direct him to your art department. Might you be of assistance? | 04:32 |
| === me_ is now known as alphainu | ||
| Keybuk | sadmac: if we want to be better than a Mac, first we have to catch up with them ;) | 05:37 |
| alphainu | ~_~ | 05:38 |
| ion | I *wish* Ubuntu looked like MacOSX (except for the even more blurry fonts and a dock that takes up a huge amount of screen real estate). :-P They have very stylish things going on, such as no left/bottom/right borders in windows, as the shadows are enough to differentiate a window of a certain color from another similar one. I love that. | 13:56 |
| sadmac | much as I liked making brown jokes in the past, I don't much care what ubuntu looks like in the end. | 16:32 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!