Alright, I have updated version 5.1 to version 5.6 without uninstalling the old one by using update-alternatives. This works just fine on our production/master server, but our replication slave seems to be having trouble.
First, in the processlist mysqld_safe just keeps starting the logger tool indefinitely until hundreds of instances are running and I restart the server manually.
Secondly, the server only runs for a couple seconds at a time without actually restarting. Whenever I run status in the MySQL shell, it has to reconnect and the status print says something like Uptime: 9 sec, even if it's been running for hours.
Output from SHOW FULL PROCESSLIST:
+----+-------------+-----------+------+---------+------+-----------------------------------------+-----------------------+
| Id | User | Host | db | Command | Time | State | Info |
+----+-------------+-----------+------+---------+------+-----------------------------------------+-----------------------+
| 1 | system user | | NULL | Connect | 0 | Checking master version | NULL |
| 2 | system user | | NULL | Connect | 0 | Waiting for the next event in relay log | NULL |
| 3 | root | localhost | NULL | Query | 0 | init | SHOW FULL PROCESSLIST |
+----+-------------+-----------+------+---------+------+-----------------------------------------+-----------------------+
The syslog seems fine, except that it complains about every table in the performance_schema database like this:
[ERROR] Native table 'performance_schema'.'session_account_connect_attrs' has the wrong structure
What's going wrong and how do I fix it?
