minimizecreation of connections during the life of the application and minimize pool
resizing.Use a lower number if the pool load is expected to be small, to minimize resource
consumption.
■Maximumconnections
Themaximum number of connections that a pool can have at any given time. Defaults to 32.
Usethis setting to enforce a limit in the amount of connection resources that a pool or
applicationcan have. This limit is also benecial to avoid application failures due to
excessiveresource consumption.
■Idletimeout
Themaximum amount in seconds that a connection is ensured to remain unused in the
pool.After the idle timeout, connections are automatically closed. If necessary, new
connectionsare created up to the minimum number of connections to replace the closed
connection.Note that this setting does not control connection timeouts enforced at the
databaseserver side. Defaults to 60 seconds.
Settingthis attribute to –1 prevents the connections from being closed. This setting is good
forpools that expect continuous high demand. Otherwise, keep this timeout shorter than
thedatabase server-side timeout (if such timeouts are congured on the specic vendor
database),to prevent accumulation of unusable connections in the pool.
■Waittimeout
Theamount of time in seconds that a request waits for a connection in the queue before
timingout. After this timeout, the user sees an error. Defaults to 60.
Settingthis attribute to 0 causes a request for a connection to wait indenitely. This setting
couldalso improve performance by keeping the pool from having to account for connection
timers.
■Validationmethod
Themethod used by the pool to determine the health of a connections in the pool. Defaults
too.
Ifa validation method is used, the pool executes a sanity check on a connection before
leasingit to an application. The eectivity and performance impact depends on the method
selected:
■meta-datais less expensive than table in terms of performance, but usually less
eectiveas most drivers cache the result and do not use the connection, providing false
results.
■tableis almost always eective, as it forces the driver to perform an SQL call to the
database,but it is also the most costly.
■auto-commitcan provide the best balance of eectiveness and performance cost, but a
numberof drivers also cache the results of this method.
■ValidationTable Name
UsingMonitoringData to TuneYour Server
SunJavaSystem Web Server 7.0 Update 1 Performance Tuning,Sizing, and Scaling Guide •76