Comment 5 for bug 1676565

We've been testing on 67 as well as 70, 71, 72, and have only seen one instance of this issue on 67 so far. The last kernel to work without issue was 4.4.0-57.

I apologize for the lack of actual evidence here, but there's really not much going on when the issue occurs - a single apparmor_parser thread might be running, or our code might have stacked up 10 or more of them trying to reload profiles. There's not even a consistent failure mode for aa-status - it will sometimes continue to return, will sometimes pause for a seconds to minutes before returning, and sometimes immediately enter a disk-sleep state and never respond.

Any suggestions for further data collection?