Often times it is possible to do a /etc/init.d/mysql start, and get a report back of a green [OK], only later to find out that mysql actually quit with an error immediately after restarting...requiring you to zap the process ID, and parse the mysqld.err to find the error. The /etc/init.d/mysql really needs to be able to monitor these error conditions and be able to tell you if mysql actually did start properly. Suggestions on how to do this are open.
I'm tagging along with an existing bug, I'll make a new one if need be though: I did /etc/init.d/net.eth0 stop and /etc/init.d/mysqld stop then /etc/init.d/mysqld start... and my ethernet came back up!!! WTH?!!! mysql doesn't need the net! I wanted to use it without starting external links.
Steven: These cases are extremely difficult to catch, as mysql reports success back to us, even if you use the mysql-provided startup scripts (eg mysql.server), and it only fails a little bit after that. there are also a few cases where it doesn't even write anything to the error log, so you can't rely on that. All of the cases where it does happen indicate that mysql is not correctly configured on the user side. If you put together a patch that catchs this, I'll be glad to include it, and see that it's sent upstream. Otherwide, the return on the investment of writing it myself is not sufficent for it to be worthwhile. aaron: that issue is fixed now.