Shell脚本中的Java DB连接池

Question

Does connection pooling db connections make sense when java code is called from a shell script, or is it better to use individual connections? Doesn't the jvm exit after every call to the shell script, forcing the db pool/factory/etc to be re-created each time the script is called?

For example, I have an external process that makes calls to a shell script, this shell script then calls a java class which executes 1 or more DB operations (query, insert, update, delete.), depending on the invoked operation. This is a standalone batch process which does not use a servlet container like tomcat. This shell script gets called by my external process over and over.

My environment is a bash shell script calling a java application (main() method) which utilizes spring jdbc and commons-dbcp 1.4 for it db processing and pooling.

Thanks in advance!

Answer 1

You should think about using a Connection Pool if you have so many calls to the database, that initializing Connections becomes costly.

Otherwise, for a batch process for which (or if) you can serialize the running tasks, one Connection would be sufficient. Spring-JDBC also provides a single connection "pool" factory object for testing, which should work for you as well (so you don't have to deal with the boiler-plating).

Answer 2

In your case, a connection pool probably only has a limited benefit, since you rightly said that whenever the JVM is shut down and started again, you would have to create the connection pool again. You could still use the pool if you want to benefit from using the DataSource interface instead of creating the connection directly. If your processing is done in a single thread, make sure that the size of the connection pool is only 1.

If you're using multiple threads in the same process, then the connection pool probably makes more sense and you could benefit from its features.

Answer 3

1. Opening a database connection is very expensive affair both in terms of resources and performance (response time).

2. When you don't use connection pool , you obtain the connection, use it and then close it. This is repeated every time whenever a DB interaction is needed.

3. A connection pool saves your from this by creating a pool of connection which can be reused, thus saving the time and resources of creating connection every time.

4. It does not matter whether you use servlet container or standalone application . It depends on the way your application (web/standalone) is interacting the database.

5. If your application has lot of parallel operation involving concurrent database access, using connection pool will be very useful.

6. So the choice and relevance of connection pool depend on the way application interacts with DB, irrespective of whether its web application or standalone.