| JanC | ducktype: my guess is that upstart thinks mysql is started before it's actually listening | 00:38 |
|---|---|---|
| ducktype | oh good | 00:41 |
| ducktype | checking that one | 00:41 |
| JanC | at least, that could explain things ;) | 00:42 |
| ducktype | a simple pre-start sleep should do the trick | 00:42 |
| ducktype | nothing :( | 00:44 |
| ducktype | mm wait | 00:44 |
| JanC | you can try to log something before it starts mydns | 00:45 |
| JanC | to make sure it actually tries to start it | 00:45 |
| ducktype | mm why: post-start script sleep 3 end script | 00:48 |
| ducktype | in the mysql job make start mysql wait forever, like when i specified "task" before to try | 00:49 |
| ducktype | JanC: i've logged from mysql pre/post start and mydns pre/post start i nto a file | 00:55 |
| ducktype | i only see one line mysql pre-start end | 00:56 |
| ducktype | seems thath the mysql job is't considered finished | 00:56 |
| JanC | so basically upstart doesn't detect when mysql is finished starting | 00:56 |
| ducktype | hangs on exec /usr/sbin/mysqld | 00:56 |
| JanC | you might need an "expect" stanza | 00:57 |
| JanC | or cheat and start mysqld from pre-start instead ;) | 00:58 |
| JanC | hm, it seems like Ubuntu starts mysqld without such tricks... | 00:59 |
| JanC | ducktype: this is the default mysql.conf in Maverick: http://paste.ubuntu.com/552322/ | 01:01 |
| ducktype | thamk you, checking.. | 01:02 |
| JanC | things like the AppArmor profile you probably don't need | 01:02 |
| ducktype | i'm using mariadb not mysql but they are really the same | 01:15 |
| ducktype | seems to be: exec /usr/sbin/mysqld & | 01:15 |
| ducktype | the & | 01:15 |
| ducktype | now doesn't hang on exec | 01:15 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!