I tried it with a simplified test case with just DELETEing a lot of data from the table with 1.5mil rows.
The DELETE command took like a minute and a half to return. After it returned, I immediately did 'stop mysql' and saw the KILL message in the daemon.log after ~5 seconds. After I started mysql again, I got the recovery messages in /var/log/mysql/error.log like this: InnoDB: Log scan progressed past the checkpoint lsn 0 241337601 101205 23:35:25 InnoDB: Database was not shut down normally! InnoDB: Starting crash recovery. InnoDB: Reading tablespace information from the .ibd files... InnoDB: Restoring possible half-written data pages from the doublewrite InnoDB: buffer... With the mysql package from -proposed (5.1.41-3ubuntu12.8) when I issued 'stop mysql' after DELETE, it was just sitting there, nothing in daemon.log, and in mysql/error.log I got 101205 23:51:25 [Note] Event Scheduler: Purging the queue. 0 events 101205 23:51:27 InnoDB: Starting shutdown... 101205 23:51:40 InnoDB: Shutdown completed; log sequence number 0 258830313 101205 23:51:40 [Note] /usr/sbin/mysqld: Shutdown complete so it was waiting for InnoDB to finish whatever it was doing (that took 13 seconds) and then it stopped. After starting mysql up again there was obviously no recovery needed. Seems to me the fix is working fine. -- You received this bug notification because you are a member of Ubuntu Server Team, which is subscribed to mysql-dfsg-5.1 in ubuntu. https://bugs.launchpad.net/bugs/620441 Title: MySQL upstart stop job does not cleanly shutdown mysql -- Ubuntu-server-bugs mailing list Ubuntu-server-bugs@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-server-bugs
