MySQL / MariaDB shuts down after exactly 1 minute 42 seconds

Question

I've installed Homestead via VirtualBox and Vagrant on macOS Sierra. Everything works, except for some mysterious reason MySQL shuts itself down after the box has been up for longer than 1 minute 42 seconds...

I'm able to use MySQL without issue for the first 1 minute 42 seconds, but then suddenly the service shuts down and refuses further connections.

I tailed /var/log/syslog after the server came up, and the following syslog entries appearing:

Dec 10 20:02:25 homestead systemd[1]: mariadb.service: Start operation timed out. Terminating.
Dec 10 20:02:25 homestead mysqld[1611]: 2016-12-10 20:02:25 140574075632384 [Note] /usr/sbin/mysqld: Normal shutdown
Dec 10 20:02:25 homestead mysqld[1611]: 2016-12-10 20:02:25 140574075632384 [Note] Event Scheduler: Purging the queue. 0 events
Dec 10 20:02:25 homestead kernel: [  100.520787] audit_printk_skb: 15 callbacks suppressed
Dec 10 20:02:25 homestead kernel: [  100.520844] audit: type=1400 audit(1481400145.309:17): apparmor="DENIED" operation="sendmsg" info="Failed name lookup - disconnected path" error=-13 profile="/usr/sbin/mysqld" name="run/systemd/notify" pid=1611 comm="mysqld" requested_mask="w" denied_mask="w" fsuid=113 ouid=0
Dec 10 20:02:25 homestead mysqld[1611]: 2016-12-10 20:02:25 140573193234176 [Note] InnoDB: FTS optimize thread exiting.
Dec 10 20:02:25 homestead mysqld[1611]: 2016-12-10 20:02:25 140574075632384 [Note] InnoDB: Starting shutdown...
Dec 10 20:02:25 homestead systemd[1]: Started Session 4 of user vagrant.
Dec 10 20:02:26 homestead mysqld[1611]: 2016-12-10 20:02:26 140574075632384 [Note] InnoDB: Waiting for page_cleaner to finish flushing of buffer pool
Dec 10 20:02:27 homestead mysqld[1611]: 2016-12-10 20:02:27 140574075632384 [Note] InnoDB: Shutdown completed; log sequence number 1355249813
Dec 10 20:02:27 homestead mysqld[1611]: 2016-12-10 20:02:27 140574075632384 [Note] /usr/sbin/mysqld: Shutdown complete
Dec 10 20:02:27 homestead kernel: [  102.937893] audit: type=1400 audit(1481400147.726:18): apparmor="DENIED" operation="sendmsg" info="Failed name lookup - disconnected path" error=-13 profile="/usr/sbin/mysqld" name="run/systemd/notify" pid=2030 comm="mysqld" requested_mask="w" denied_mask="w" fsuid=113 ouid=0
Dec 10 20:02:27 homestead kernel: [  102.938561] audit: type=1400 audit(1481400147.726:19): apparmor="DENIED" operation="sendmsg" info="Failed name lookup - disconnected path" error=-13 profile="/usr/sbin/mysqld" name="run/systemd/notify" pid=1611 comm="mysqld" requested_mask="w" denied_mask="w" fsuid=113 ouid=0
Dec 10 20:02:27 homestead systemd[1]: Failed to start MariaDB database server.
Dec 10 20:02:27 homestead systemd[1]: mariadb.service: Unit entered failed state.
Dec 10 20:02:27 homestead systemd[1]: mariadb.service: Failed with result 'timeout'.
Dec 10 20:02:27 homestead systemd[1]: Reached target Multi-User System.
Dec 10 20:02:27 homestead systemd[1]: Starting Notify bootloader that boot was successful...
Dec 10 20:02:27 homestead systemd[1]: Reached target Graphical Interface.
Dec 10 20:02:27 homestead systemd[1]: Starting Update UTMP about System Runlevel Changes...
Dec 10 20:02:27 homestead systemd[1]: Started Update UTMP about System Runlevel Changes.
Dec 10 20:02:27 homestead snap[2063]: Ignoring 'booted' on classic
Dec 10 20:02:27 homestead systemd[1]: Started Notify bootloader that boot was successful.
Dec 10 20:02:27 homestead systemd[1]: Startup finished in 5.989s (kernel) + 1min 36.860s (userspace) = 1min 42.850s.

The first line, where it says mariadb.service, gets logged the moment I get kicked out of MySQL. But I have no idea why.

It looks like the logs are implying MySQL / MariaDB are still trying to start up after the box has been brought up - even though I've been able to connect to them by this point without issue. Then some mystery timeout is hit, and the whole thing shuts down.

Can anyone help me interpret this...? Thanks!

Answer 1

Finally figured it out - AppArmor seems to be causing the problem.

By following the steps outlined here, I was able to resolve the issue, and MySQL no longer shuts down unexpected.