| Keybuk | no idea there | 00:32 |
|---|---|---|
| Keybuk | I can look | 00:32 |
| Keybuk | ion: what are you subscribed as? | 00:37 |
| Keybuk | oh, wait, wrong ML | 00:37 |
| Keybuk | I can't look :p | 00:37 |
| === Keybuk changed the topic of #upstart to: Upstart 1.1 "It's probably 'cause you think you're cooler than me" | http://upstart.at/ | ||
| telive | hello everyone . i have just installed apache2 in ubuntu and found that there is no apache2.conf file in /etc/init dir . so apache do not stat at the boot time . | 01:16 |
| Keybuk | telive: it may have a shell script in /etc/init.d | 01:16 |
| telive | Keybuk, yes | 01:16 |
| Keybuk | there you go then | 01:16 |
| telive | Keybuk, but the apache2 didn't start either . | 01:17 |
| Keybuk | that I can't answer, you would need to talk to Ubuntu people | 01:17 |
| telive | Keybuk, i have checked the status of apache2 in sysv-rc-conf , it is on | 01:17 |
| Keybuk | I'd suggest #ubuntu-server would be a good start, should such a channel exist | 01:18 |
| telive | can anyone send me your apache2.conf file in /etc/init ? | 01:18 |
| telive | Keybuk, thanks | 01:19 |
| JanC | telive: did you check logs for any error message or such? | 01:51 |
| JanC | nevermind, I see you're in -server already ☺ | 01:52 |
| telive | JanC, hehe | 01:53 |
| telive | JanC, as i know , upstart have take place of sysv in ubuntu . and there is no apache2.conf file in /etc/init , so the apache server didn't startup | 01:54 |
| JanC | upstart in Ubuntu has sysv emulation, so the sysv init scripts in /etc/init.d/ should work | 01:56 |
| === apachelogger is now known as apachelogger_ | ||
| === apachelogger_ is now known as apachelogger__ | ||
| === apachelogger__ is now known as apachelogger___ | ||
| === apachelogger___ is now known as apachelogger | ||
| wraiden | hello, fetching http://upstart.at/download/1.x/upstart-1.1.tar.gz gives a 403 Forbidden response... | 19:19 |
| Hobart | is there a way to ask upstart to prompt you for each step it's going to take (a-la the old "interactive startup") to troubleshoot while booting? | 20:54 |
| marrusl | Hobart, I don't think so. but take a look here: http://upstart.ubuntu.com/wiki/Debugging | 21:43 |
| marrusl | Hobart, the main takeaway being adding --verbose to the kernel command line. there's also --debug which goes further. | 21:43 |
| JanC | during the first steps there probably would be no way yet to prompt? | 21:43 |
| marrusl | and then looking at daemon.log to see what happened. | 21:44 |
| marrusl | JanC, because you have multiple paths going at once, i'm not sure an interactive boot even makes sense. | 21:46 |
| JanC | well, it would remove a lot of the parallelism | 21:52 |
| JanC | and hide most of the possible race conditions | 21:52 |
| marrusl | JanC, indeed. and then you wouldn't be troubleshooting your actual scenario. | 21:52 |
| marrusl | yup exactly. | 21:52 |
| Hobart | marrusl -> thank you for the information! | 22:50 |
| Hobart | unfortunately at a certain step in bootup, the system was cold-booting | 22:51 |
| Hobart | which is why I was hoping for a [Y/n]? prompted option | 22:51 |
| Hobart | might open up a wish request item for it | 22:51 |
| marrusl | I've locked up a boot before daemon.log even gets started, thankfully I was able to reverse the change. but yeah, that sounds tricky then. | 22:52 |
| marrusl | Hobart, does it cold boot when you try to boot to single user? | 22:52 |
| Hobart | marrusl -> it wasn't, but I've already made a few changes since then (I'm 95% sure I have unstable RAM and am awaiting ECC replacement modules) | 22:53 |
| Hobart | I -think- the issue was that I'd said for the onboard radeon 3100 integrated motherboard gfx adapter to grab 1GB instead of 'Auto' (256MB) which appears to be outside the stable test cases for both the Windows and Linux drivers :-( (flashplayer can BSoD the Windows side in that setting, and X was coldbooting / hardlocking, no alt-sysrq-R sanity recovery) | 22:54 |
| Hobart | which leads me to believe that the stupid was centered in the DRM module or other kernel level thing, since just X chokng usually lets me switch out virtual consoles or deal with X's recovery | 22:55 |
| marrusl | Hobart, aha. Just thinking you could start up services manually from runlevel 1 or even see if going to runlevel 2 manually makes a difference. | 22:56 |
| marrusl | Hobart, ah, but if it hardware, then may as well wait for the parts and see. good luck! | 22:56 |
| Hobart | thanks :) I still think 'boot everything serially' will make a good wishlist item for troubleshooting : | 22:58 |
| Hobart | :) | 22:58 |
| JanC | Hobart: the way upstart works makes it *very* difficult to determine what a serial boot should do ;) | 23:37 |
| Hobart | JanC -> Imagine that I launch upstart under gdb with a breakpoint set on the call to execve() ... | 23:48 |
| JanC | but the order could still be different every time, and it would both hide bugs & create new problems if you do that... | 23:51 |
| JanC | not to mention that it's very well possible that at some point there is no way to tell upstart to continue and you'd get stuck... | 23:54 |
| Hobart | Putting a fire axe on the wall of an apartment building isn't assurance that you'll get out alive. | 23:56 |
| Hobart | It's still nice to have there. :) | 23:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!