> InnoDB: Error: semaphore wait has lasted > 600 seconds
> InnoDB: We intentionally crash the server, because it appears to be hung.
What was the server doing when it crashed?
Several such reports at
https://bugs.mysql.com/bug.php?id=73890, none of them quite solved, several ideas to try.
It's still memory stress.
Have you tried running top and free, say every hour or even more frequently?
Please study Jeremy Cole's piece on MySQL and memory pressure at
https://blog.jcole.us/2012/04/16/a-brief-update-on-numa-and-mysql/, wher ehe described the solution that works with Twitter MySQL.
Have you run iostat or pt-diskstats?
Is there any swapping?