| danolj | morning | 17:53 |
|---|---|---|
| danolj | so how do I make it so that respawn doesn't go into effect after a stop has been issued? | 18:00 |
| JanC | it should never do that AFAIK | 18:01 |
| JanC | danolj: maybe somehow it crashes? | 18:02 |
| danolj | http://pastebin.com/xu9P54ig | 18:03 |
| danolj | that's the upstart script for our redis instance | 18:03 |
| danolj | in the pre-stop script the CLI line tells the server to shutdown gracefully | 18:03 |
| danolj | in the redis logs it is indeed shutting down as expected, however, upstart/init simply starts things up again | 18:03 |
| danolj | exit status problem perhaps? | 18:05 |
| JanC | maybe, although I thought that doesn't matter in case you use 'stop' | 18:08 |
| === Md_ is now known as Md | ||
| danolj | JanC: yes, that is what the docs say, but I've yet to figure out why it continues to restart after stop | 18:44 |
| danolj | okay, here is what is happening | 18:46 |
| danolj | by asking redis to shutdown gracefully, init sees that it has "died" and restarts it | 18:46 |
| danolj | so if I remove the 'cli shutdown' command, all is good, however, it sure would be nice to stop things gracefully | 18:47 |
| JanC | danolj: you can tell upstart what other exit statuses / signals are allowed to stop it? | 19:30 |
| danolj | JanC: not sure I understand what you're saying | 20:29 |
| JanC | danolj: I was referring to the "normal exit" stanza | 20:31 |
| danolj | okay, let me go read about it | 20:31 |
| danolj | JanC: thank you for that pointer, allowing 'TERM' as a valid exit status fixed the problem | 21:01 |
| danolj | er, not problem, rather provided a reasonable solution | 21:02 |
| danolj | Thank you | 21:02 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!