The JOC Cockpit log states (timestamps in UTC+2):.
However, consider the problem for connection id 189 from the below logs: first MySQL aborts the connection, then the connection pool tries to re-use it:
Therefore the default settings of the connection pool to close/renew connections earlier than the DMBS seems to be fine.
However, these idle timeout settings apply only if a logical "close" connection operation is performed by the JOC Cockpit towards the connection pool. The hibernate configuration for JOC Cockpit by default includes the settings maxIdleTime=14400 and maxConnectionAge=14400 which corresponds to 4 hours.
The JOC Cockpit connection pool tries to re-use the connection which results in an error.
When aborting an idle connection then the respective MySQL message is: "Aborted connection 189 to db: 'scheduler112' user: 'scheduler' host: 'apmacwin' (Got an error reading communication packets)", see linked article for details.
MySQL in the below example is configured to use wait_timeout=28800, i.e.
After some timeout the DBMS will abort a connection.
if a user does not log out and a session timeout occurs, a connection to the database that has been created for the session seems not to be closed.
JOC Cockpit makes use of a connection pool which translates to the fact that operations such as "open" and "close" connection do not open a physical connection, but use a logical connection from the connection pool.
0 kommentar(er)
