| c9er | hey... i am trying to make an upstart script for rabbitmq but it blocks it starts the service and when i stop it the service is stopped but the stop process hangs and if killed it and try to start it again the start hangs | 08:34 |
|---|---|---|
| jhunt_ | c9er: sounds like Upstart is unable to find the pid of the rabbitmq process. Have you specified "expect fork" or "expect daemon"? http://upstart.ubuntu.com/cookbook/#expect | 08:36 |
| c9er | yes i did | 08:37 |
| jhunt_ | which one? maybe you specified the wrong option (they are different) | 08:37 |
| c9er | expect fork | 08:37 |
| jhunt_ | I don't know rabbitmq, but if it forks twice, that won't work. | 08:38 |
| c9er | what if it forks more than twice? | 08:38 |
| jhunt_ | it wouldn't need to. 2 forks is the maximum number of forks any daemon needs to perform to initialize. | 08:39 |
| c9er | will then i guess i will try the expect daemon | 08:39 |
| jhunt_ | it might subsequently fork off process of course to handle incoming requests, but we're talking solely about daemon initialisation this point. | 08:39 |
| jhunt_ | That *should* work, but if it doesn't, you can cheat: http://upstart.ubuntu.com/cookbook/#alternative-method | 08:40 |
| c9er | another question is there anyway to make the upstart reread the files without rebooting | 08:40 |
| c9er | because you know i define an instance then i remove it then i stop ther service but the upstart insists that i have to define that instance | 08:41 |
| jhunt_ | there is a caveat though - although that will ensure the process starts, upstart still won't know the pid, so won't be able to stop it gracefully. | 08:41 |
| jhunt_ | you want upstart to re-read its .conf files? | 08:42 |
| jhunt_ | removing a .conf file when an instance is running, is going to cause you problems :) | 08:42 |
| jhunt_ | upstart re-reads its .conf files automatically whenever any of them changes. | 08:43 |
| c9er | no they don't | 08:43 |
| jhunt_ | ? | 08:43 |
| c9er | and i mean by instance the instance variable inside the file | 08:43 |
| c9er | it keeps telling me that variable need a value | 08:44 |
| c9er | while i have already removed it's line from he file | 08:44 |
| jhunt_ | If you have a running job that needs an instance variable, yes, you'll need to specify the value of that instance variable to stop that running instance. | 08:45 |
| jhunt_ | See http://upstart.ubuntu.com/cookbook/#instance | 08:45 |
| jhunt_ | Also see: http://upstart.ubuntu.com/cookbook/#attempting-to-stop-a-job-that-requires-an-instance-variable | 08:45 |
| c9er | i don't mean so | 08:45 |
| c9er | i have already removed the line of the variable after stopping the service | 08:46 |
| c9er | the next time i need to start it it asked for the value | 08:46 |
| jhunt_ | c9er: sounds like you might have another instance of that job still running? | 08:46 |
| c9er | nope i check the ps output each time | 08:47 |
| c9er | i think i may report it as a bug | 08:47 |
| jhunt_ | please do, ensuring you specify exactly how to reproduce the problem. | 08:49 |
| jhunt_ | Going back to what you say above: "i define an instance then i remove it then i stop ther service" | 08:49 |
| jhunt_ | this is wrong - you have to start the instance, stop then instance *then* remove the variable from the file. | 08:49 |
| c9er | that's exactly what i have done | 08:54 |
| c9er | :) | 08:54 |
| c9er | i have wrote upstart script on ubuntu 10.10 version 6.6 it'w working perfectly but when i use the same script on upstart 9.7.3 on ubuntu 11.04 it freezes | 11:39 |
| c9er | any help?! | 11:42 |
| jhunt_ | c9er: please raise a bug with the details: https://bugs.launchpad.net/ubuntu/+source/upstart/+filebug | 11:57 |
| c9er | any idea why would the upstart script hang when i use stop? | 16:42 |
| c9er | kay does upstart have any kind of a data base locking file or caching thing? | 17:02 |
| JanC | c9er: AFAIK upstart keeps all its "state" data in RAM, and the only thing resembling a database is the collection of *.conf files? | 17:30 |
| c9er | JanC, is there anyway to access these data in RAM? | 17:35 |
| JanC | not really | 17:37 |
| JanC | you can get some of it using some 'initctl' commands like "list" | 17:39 |
| c9er | well do you know why would stop hangs? | 17:41 |
| JanC | and of course you can turn up log verbosity & check the log messages | 17:41 |
| c9er | JanC, how is that? | 17:41 |
| JanC | c9er: not without context | 17:41 |
| JanC | c9er: read "man initctl" for the command to change log verbosity | 17:42 |
| c9er | k | 17:42 |
| JanC | it's named 'log-priority' | 17:42 |
| JanC | you can also set it on the kernel commandline in Ubuntu (see the cookbook) | 17:43 |
| c9er | JanC, so any idea why would the script hang on stop? | 17:44 |
| JanC | c9er: script? | 17:44 |
| JanC | c9er: did you read http://upstart.ubuntu.com/cookbook/#debugging ? | 17:45 |
| c9er | yea i did | 17:46 |
| [dlp] | Enhancement request: | 21:50 |
| [dlp] | At boot time, with "noisy" output (and no splash screen). | 21:50 |
| [dlp] | .. | 21:50 |
| [dlp] | Information about what process is being started left justified... | 21:51 |
| [dlp] | Indication of success: right justified. | 21:51 |
| [dlp] | Most monitors are wide enough to make it impossible to correlate one to the other. | 21:51 |
| [dlp] | Either print them together and be jaggedy or pad it with dots. | 21:52 |
| [dlp] | If anyone agrees and is prepared to point me inthe direction of exactly the right source file I'll do it myself. | 22:03 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!