> The scenario is this - the connection to the database goes down, CAS tries to
> connect and fails. Connection to database comes back up. CAS continues to
> fail until it is restarted.
Assuming a connection pool, which is a best practice, there are two
solutions: connection validation on idle and connection validation on
checkout. Both have performance costs and availability benefits, but
you should know what high-level availability behavior you want before
you go turning knobs. If you want CAS to come back _eventually_, then
idle testing is arguably the best solution. In that case you need the
following property mentioned in the JpaTicketRegistry [1] docs:
p:idleConnectionTestPeriod="${database.pool.idleConnectionTestPeriod}"
You can read more about various c3p0 connection validation options [2]
if you're curious about other types of validation.
M
[1] https://wiki.jasig.org/display/CASUM/JpaTicketRegistry
[2] http://www.mchange.com/projects/c3p0/#configuring_connection_testing
--
You are currently subscribed to [email protected] as:
[email protected]
To unsubscribe, change settings or access archives, see
http://www.ja-sig.org/wiki/display/JSG/cas-user
