Comment 8 for bug 1621852

Thanks for the input.

The server isn't swapping - no traces of swap anywhere.

The gdb capture was taken once the problem happened again, so that is the state of the server at the time replication wasn't going ahead.

Again, the only solution (which prevents this from happening) is to convert the table to InnoDB.
I have pinged MariaDB in the ticket you found so they can check the stacktraces too.

Is there anything else I can do to narrow this problem a bit more so we can find out what's going on or get more troubleshooting scenarios?