How to reduce a batch job's execution time?

[Hallgeir]

I have problems with the execution speed for a new batch job where I’m using the spring batch framework.

The job are reading and committing 10 customers at a time.

From the log, it looks like the reading from the database is very slow. From the mapRow method is first called, the customers are processed (the process method is called), and the customers are written to the database (the write method is called) takes less than one second. But from the execution of the write is completed to the next time mapRow is called, takes approximately 10 seconds each time.

Does anyone have experience with how a job can be configured to reduce the execution time?

@Configuration
public class CustomerAndAccountReaderConfiguration {

    @Bean
    public ItemReader<CustomerAndAccountOldDTO> customerAndAccountReader(DataSource dataSource,
            PagingQueryProvider pagingQueryProvider, PersonDAO personDAO, AccountDAO accountDAO,
            AddressDAO addressDAO) throws Exception {
        JdbcPagingItemReader<CustomerAndAccountDTO> jdbcPagingItemReader = new JdbcPagingItemReader<CustomerAndAccountDTO>();
        jdbcPagingItemReader.setFetchSize(10);
        jdbcPagingItemReader.setDataSource(dataSource);
        jdbcPagingItemReader.setQueryProvider(pagingQueryProvider);
        jdbcPagingItemReader.setRowMapper(new CustomerRowMapper(personDAO, accountDAO, addressDAO));
        jdbcPagingItemReader.afterPropertiesSet();
        return jdbcPagingItemReader;
    }

    @Bean
    public PagingQueryProvider pagingQueryProvider() {
        Db2PagingQueryProvider db2PagingQueryProvider = new Db2PagingQueryProvider();
        db2PagingQueryProvider.setSelectClause("select * ");
        db2PagingQueryProvider.setFromClause("from Customer ");
        db2PagingQueryProvider.setSortKey("id");
        return db2PagingQueryProvider;
    }

}



<beans>
    <batch:job id="databaseconversion">
        <batch:step id="convertcustomerandaccount">
            <batch:tasklet>
                <batch:chunk 
                    reader="customerAndAccountReader"
                    processor="customerAndAccountProcessor" 
                    writer="customerAndAccountWriter"
                    commit-interval="10" 
                />
            </batch:tasklet>
        </batch:step>
    </batch:job>
</beans>