our application is using Spring for TX management and is marking certain transactions as readonly. When deploying our application on websphere (8.5.5.3) with a Oracle JDBC Connection we are getting exceptions like the following:
Caused by: java.sql.SQLException: DSRA9010E: 'setReadOnly' is not supported on the WebSphere java.sql.Connection implementation.
at com.ibm.ws.rsadapter.spi.InternalOracleDataStoreHelper.setReadOnly(InternalOracleDataStoreHelper.java:371)
at com.ibm.ws.rsadapter.jdbc.WSJdbcConnection.setReadOnly(WSJdbcConnection.java:3646)
at org.springframework.jdbc.datasource.LazyConnectionDataSourceProxy$LazyConnectionInvocationHandler.getTargetConnection(LazyConnectionDataSourceProxy.java:410)
at org.springframework.jdbc.datasource.LazyConnectionDataSourceProxy$LazyConnectionInvocationHandler.invoke(LazyConnectionDataSourceProxy.java:376)
at com.sun.proxy.$Proxy476.getMetaData(Unknown Source)
at sun.reflect.GeneratedMethodAccessor40.invoke(Unknown Source)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:55)
at java.lang.reflect.Method.invoke(Method.java:619)
at org.springframework.jdbc.datasource.TransactionAwareDataSourceProxy$TransactionAwareInvocationHandler.invoke(TransactionAwareDataSourceProxy.java:239)
at com.sun.proxy.$Proxy476.getMetaData(Unknown Source)
I know what websphere is trying to tell me but I was wondering if there is a way to disable this check so that Connection.setReadonly calls are just ignored instead of throwing an exception.
Of course I could also change the application to not use readonly transactions but that would be much more complicated.
Try to unwrap the connection object like this:
Context ic = new InitialContext();
DataSource ds = (DataSource)ic.lookup("jdbc/OracleDS");
Connection conn = ds.getConnection();
if (conn.isWrapperFor(oracle.jdbc.OracleConnection.class)) {
// Returns an object that implements the given interface to
// allow access to non-standard methods, or standard methods
// not exposed by the proxy.
oracle.jdbc.OracleConnection oraCon = conn.unwrap(oracle.jdbc.OracleConnection.class);
// Do some Oracle-specific work here.
oraCon.setReadOnly(readOnly);
....
}
conn.close();
See WebSphere Application Server and the JDBC 4.0 Wrapper Pattern
Oracle JDBC (12c onwards; perhaps 11g as well?) can be tricky to use with readonly -
according to https://marschall.github.io/2017/01/28/oracle-read-only.html:
Calling Connection.setReadOnly(true) with the 12c driver no longer establishes a read only transaction
This means that Spring (before version 4.3.7) also struggles to set read only transactions with Oracle JDBC (see previous link).
To overcome this, you need to manually include SET TRANSACTION READ ONLY in your SQL, instead of relying on Spring's #Transactional(readOnly=true).
However, from Spring 4.3.7, the transaction now acts correctly, and so you should no longer see this issue (https://github.com/spring-projects/spring-framework/issues/19774)
Related
I am setting up a Spring-boot application to connect to HP NonStop Tandem's SQL/MX. First I achieved this connection by hard-coding the jdbc parameters like dataSource, URL, etc in the service section of the application and it worked (I was able to access tables by executing query).
Now I am trying to remove the hard coded part and have my database related info in application.properties file, but now I am getting the following error
org.springframework.jdbc.support.MetaDataAccessException: JDBC DatabaseMetaData method not implemented by JDBC driver - upgrade your driver; nested exception is java.lang.AbstractMethodError: Method com/tandem/t4jdbc/SQLMXConnection.isValid(I)Z is abstract
Can someone help me understand the root cause? The same driver jar is being used when hard-coding the datasource details and it worked but not working when having the data source properties in application.properties and needs an upgrade to the jar.
I encountered the same exception when using Spring Data JPA in a Spring Boot application, the JTDS driver and the Hikari connection pool. In my case I discovered that the following fixed the problem:
Examining the class com.zaxxer.hikari.pool.PoolBase, the following can be observed:
this.isUseJdbc4Validation = config.getConnectionTestQuery() == null;
Thus JDBC 4 validation will not be attempted if there is a connection test query configured. In a Spring Boot application, this can be accomplished like this:
spring.datasource.hikari.connection-test-query=select 1;
Regretfully I do not have any experience with the T4SQLMX driver but nevertheless hope this can be of some use.
I recently fought through the same issue, for me I was using a JDBC type 3 driver; but my spring implementation only supported a type 4 driver, thus when the method you linked above was attempted to be called, it caused the error.
I suggest you look for a type 4 driver for your particular database and see if that resolves your issue.
I am working now with oracle and spring jdbc but I don't want to use the schema in my sql statements:
Example: Select * from SCHEMA.table
Is there any way to set default schema in application.properties or application.yml?
Assuming you define your database connections using spring datasources, you can set the default schema when defining the datasource configuration:
spring.datasource.schema = #value for your default schema to use in database
You can find more info here: Spring Boot Reference Guide. Appendix A. Common application properties
After doing some research, looks like Oracle driver doesn't let you set a default schema to work with, as noted here:
Default Schema in Oracle Connection URL
From that post, you have two options:
Execute this statement before executing your statements:
ALTER SESSION SET CURRENT_SCHEMA=yourSchema
Create synonyms for your tables/views/etc (which I find really cumbersome if we're talking about lots of elements in your database).
I would advice using the first option. From what I see, Spring boot doesn't offer a simple way to execute a statement when retrieving the connection, so the best bet will be to use an aspect around the getConnection method (or the method that retrieves the connection from the data source) and execute the statement there.
From your comment, an easier way to solve it is by using a script in spring.datasource.schema:
spring.datasource.schema = schema.sql
And then a file squema.sql with the following:
ALTER SESSION SET CURRENT_SCHEMA=mySchema
In spring boot, I've found another way of doing it,
#Bean
#ConfigurationProperties(prefix="spring.datasource")
public DataSource dataSource(#Value("${spring.datasource.schema}") String schema) {
DataSource datasource = DataSourceBuilder.create().build();
if(!schema.isEmpty() && datasource instanceof org.apache.tomcat.jdbc.pool.DataSource){
((org.apache.tomcat.jdbc.pool.DataSource) datasource).setInitSQL("ALTER SESSION SET CURRENT_SCHEMA=" + schema);
}
return datasource;
}
I found another way to get around this by updating entity class with
#Table(schema = "SCHEMA_NAME" ,name = "TABLE_NAME")
If you are using hikari, use spring.datasource.hikari.schema=YOUR_SCHEMA.
Works for me with SpringBoot + tomcat using Oracle.
I was having issues with the currently accepted answer; specifically, the schema would only be changed from the initial connection. If your app uses a connection pool, you need to configure the pool to apply SQL for each connection.
For instance, using the default jdbc pool in Spring Boot 1.5.x (Tomcat):
spring.datasource.tomcat.init-s-q-l = ALTER SESSION SET CURRENT_SCHEMA=mySchema
Connecting to the database as your user, you can create a trigger that will change the schema each time you login:
CREATE OR REPLACE TRIGGER LOGON_TRG
AFTER LOGON ON SCHEMA
BEGIN
EXECUTE IMMEDIATE 'ALTER SESSION SET CURRENT_SCHEMA = foo';
EXCEPTION
when others
then null;
END;
/
Another option is to create a datasource wrapper. Create the datasource as normal and then create the wrapper that forwards all methods except for the getConnection methods. For those I just added SQL to set the schema. We have multiple datasources and this allowed us to specify a different schema for each datasource. If anyone knows if there's an issue with this I'd love comments. Or if there's an alternative that uses the properties.
I have a problem that I don't know how solve and researching the net has not helped me much. I declare in glassfish 4.0 asadmin console a serializable connection pool and its corresponding resource.
create-jdbc-connection-pool --datasourceclassname oracle.jdbc.xa.client.OracleXADataSource --maxpoolsize 8 --isolationlevel serializable --restype javax.sql.XADataSource --property Password=A_DB:User=A_DB:URL="jdbc\:oracle\:thin\:#localhost\:1521\:orcl" ATestPool
create-jdbc-resource --connectionpoolid ATestPool jdbc/ATest
Then inside a stateless bean I build a datasource via jndi as follows:
InitialContext ic = new InitialContext();
jndiDataSource = (DataSource) ic.lookup("jdbc/ATest");
and I'm getting connection as follows
jndiDataSource.getConnection();
Connections are properly obtained and released via finally clauses in each method we they are needed.
However, pairing serializable connection pool with XA data sources seems not to work, as getting first connections throws the following pair of exceptions in the order shown below
JTS5041: The resource manager is doing work outside a global transaction
oracle.jdbc.xa.OracleXAException
at oracle.jdbc.xa.OracleXAResource.checkError(OracleXAResource.java:1110)
RAR5029:Unexpected exception while registering component
javax.transaction.SystemException
at com.sun.jts.jta.TransactionImpl.enlistResource(TransactionImpl.java:224)
with the following
RAR7132: Unable to enlist the resource in transaction. Returned resource to pool. Pool name: [ ATestPool ]]]
RAR5117 : Failed to obtain/create connection from connection pool [ ATestPool ]. Reason : com.sun.appserv.connectors.internal.api.PoolingException: javax.transaction.SystemException]]
RAR5114 : Error allocating connection : [Error in allocating a connection. Cause: javax.transaction.SystemException]]].
Now if the connection pool is recreated without --isolationlevel serializable, then application works fine without any changes into the code. Also, if one keeps the isolation parameter and uses non-XA transactions as
--datasourceclassname oracle.jdbc.pool.OracleDataSource
--restype javax.sql.DataSource
then again application works without any changes into the code.
I was wondering if anyone could explain to me what could be wrong in the above setup and how to actually make serializable work with XA data sources. Thanks.
I think you need to enable useNativeXA.
I'm new to JDBC. I've installed GlassFish 3.1.1 on Centos 6.2 and need to use it with an application that connects to an Oracle 11G database on another server. I've read through the documentation for GlassFish and think I understand how to create a JDBC connection pool as well as a JDBC resource. My question is, how do I use this information when coding the java middle-tier to connect to the database?
Currently (with just GlassFish install and no JDBC configuration), I am relying on the CentOS enviroment variables for java (such as CLASSPATH) to allow the web application to use the JDBC drivers. However, I'm getting the following error:
java.lang.NoClassDefFoundError: oracle/jdbc/pool/OracleDataSource
Thus, my attempt to create a JDBC connection pool and resource in GlassFish (so the app can use the JDBC driver). My java file starts out:
import java.sql.*;
import oracle.jdbc.*;
import oracle.jdbc.pool.OracleDataSource;
class JDBCexample {
public static void main(String args[]) throws SQLException {
Connection conn;
Statement stmt;
ResultSet rset;
String query;
String sqlString;
String person_firstName;
String person_lastName;
String person_email;
int person_salary;
// connect to database
OracleDataSource ds = new OracleDataSource();
ds.setURL("jdbc:oracle:thin:myID/myPWD#192.168.0.1:1521:mySID");
conn = ds.getConnection();
// read something in database
stmt = conn.createStatement();
query = "SELECT first_name, last_name, email, salary FROM HR.Employees where rownum < 6";
rset = stmt.executeQuery(query);
while (rset.next()) {
person_firstName = rset.getString("first_name");
person_lastName = rset.getString("last_name");
person_email = rset.getString("email");
person_salary = rset.getInt("salary");
System.out.format(person_firstName + " " + person_lastName + " " + person_email + " %d%n", person_salary)
}
and so on...
QUESTION: How would I change the above code after I create a JDBC Connection Pool (named: myPool) and a JDBC Resource (named: myDBPool)? If it matters, I'm using Oracle 11.2, CentOS 6.2, GlassFish 3.1.1 with mod_jk and fronted by Apache 2.2 webserver, JDK 1.6. I don't have any clustering or load-balancing.
UPDATE 1: I thought this link was a good reference (see section titled: "Creating a Data Source Instance, Registering with JNDI, and Connecting"). But when I modify the above Java file as follows (just preparing the java file; haven't touched GlassFish yet),
// Add These:
import javax.naming.Context;
import javax.naming.InitialContext;
// Change from this:
// connect to database
OracleDataSource ds = new OracleDataSource();
ds.setURL("jdbc:oracle:thin:myID/myPWD#192.168.0.1:1521:mySID");
conn = ds.getConnection();
// To this:
// connect to database
Context ctext = new InitialContext();
OracleDataSource ds = (OracleDataSource)ctext.lookup("jdbc/myDBPool");
conn = ds.getConnection();
I get the errors:
JitterClass.java:67: unreported exception javax.naming.NamingException; must be caught or declared to be thrown
Context ctext = new InitialContext();
^
JitterClass.java:68: unreported exception javax.naming.NamingException; must be caught or declared to be thrown
OracleDataSource ds = (OracleDataSource)ctext.lookup("jdbc/myDBPool");
^
UPDATE 2: I cleared those compile errors using cyril's comments below (to throw all exceptions). Then I created JDBC Connection Pool and JDBC Resource, and the Ping was successful. So then I run the application from the client and observe the following error:
java.lang.ClassCastException : com.sun.gjc.spi.jdbc40.DataSource40 cannot be cast to oracle.jdbc.pool.OracleDataSource
At this point, if I add an include javax.sql.DataSource to the program, and change this line:
OracleDataSource ds = (OracleDataSource)ctext.lookup("jdbc/myDBPool");
to become this line:
DataSource ds = (DataSource)ctext.lookup("jdbc/myDBPool");
it compiles without errors. But now I'm confused... aren't we supposed to be using OracleDataSource here? Or, does GlassFish somehow implement OracleDataSource since I do see a setting for this connection pool for Datasource Classname set to oracle.jdbc.pool.OracleDataSource (?). Hoping someone can explain this.
Do pings on your connection pool work?
If not, check your pool configuration w/ http://docs.oracle.com/cd/E18930_01/html/821-2416/beamw.html#beanh
Once pings work and the JDBC Resource is configured, you should be able to access it in your app code through JNDI:
InitialContext context = new InitialContext();
DataSource ds = (DataSource) context.lookup("jdbc/myDBPool"); // or whatever name you used when creating the resource
conn = ds.getConnection();
Hope it helps,
RESPONSE TO UPDATE 1:
That's just the compiler telling you to formally catch or declare a checked exception which may be thrown by JNDI.
For testing purposes, the easiest way out of this (and future errors like this) is to just widen your method signature to throw all exceptions, i.e.:
public static void main(String args[]) throws /*SQL*/Exception {
RESPONSE TO UPDATE 2:
There's no reason to cast JDBC interfaces down to Oracle implementations unless you need to access any custom feature not specified in the JDBC spec. The purpose of a DataSource is to be a factory for Connections, whose API is defined in the JDBC interface, so that should be all you need. When you define a connection pool and resource in GlassFish, the app server is adding value by wrapping the JDBC driver classes and proxying them seamlessly for you as long as you stick to import java.sql.*. No need for oracle imports :) The main advantage being that if you ever decide to switch to MySQL or some other data store later on, your code is then portable and doesn't need any change.
To add to cyril's good answer:
Instead of the JNDI lookup, you can also use Resource Injection to set up your DataSource:
#Resource(name = "jdbc/Your_DB_Res")
private DataSource ds;
On startup, the application server will then inject the JDBC ressource. This section of Java EE Tutorial has more on that matter.
By using resource injection, you can reduce the amount of boilerplate code. This article introduces the concepts.
Besides adding the driver to your classpath, you should try adding the appserv-rt.jar file to your project's build path (the jar is located in Glassfish's lib directory). If you don't want to include all the other jars you should first create a library containing the appserv-rt jar and then add it to your project's build path.
I try to set up a local development infrastructure and I want to use HSQLDB as a datasource with my WAS 6.1. I already know that I have to use Apache DBCP to get a connection pooling, but I'm stuck when my application tries to get the first connection.
What I've done
In WAS I created a JDBC provider with the class org.apache.commons.dbcp.cpdsadapter.DriverAdapterCPDS and removed everything from the classpath input field. Then I put commons-dbcp.jar, commons-pool.jar and hsqldb.jar in MYAPPSERVERDIRECTORY/lib/ext.
Then I created a new datasource with that provider. I added the following custom properties:
driver=org.hsqldb.jdbc.JDBCDriver
url=jdbc:hsqldb:file:///C:/mydatabase.db;shutdown=true
user=SA
password=
My Problem
When I run my application and the first connection to the database is made, I get the following exception:
---- Begin backtrace for Nested Throwables
java.sql.SQLException: No suitable driverDSRA0010E: SQL-Status = 08001, Fehlercode = 0
at java.sql.DriverManager.getConnection(DriverManager.java:592)
at java.sql.DriverManager.getConnection(DriverManager.java:196)
at org.apache.commons.dbcp.cpdsadapter.DriverAdapterCPDS.getPooledConnection(DriverAdapterCPDS.java:205)
at com.ibm.ws.rsadapter.spi.InternalGenericDataStoreHelper$1.run(InternalGenericDataStoreHelper.java:918)
at com.ibm.ws.security.util.AccessController.doPrivileged(AccessController.java:118)
at com.ibm.ws.rsadapter.spi.InternalGenericDataStoreHelper.getPooledConnection(InternalGenericDataStoreHelper.java:955)
at com.ibm.ws.rsadapter.spi.WSRdbDataSource.getPooledConnection(WSRdbDataSource.java:1437)
at com.ibm.ws.rsadapter.spi.WSManagedConnectionFactoryImpl.createManagedConnection(WSManagedConnectionFactoryImpl.java:1089)
at com.ibm.ejs.j2c.FreePool.createManagedConnectionWithMCWrapper(FreePool.java:1837)
at com.ibm.ejs.j2c.FreePool.createOrWaitForConnection(FreePool.java:1568)
at com.ibm.ejs.j2c.PoolManager.reserve(PoolManager.java:2338)
at com.ibm.ejs.j2c.ConnectionManager.allocateMCWrapper(ConnectionManager.java:909)
at com.ibm.ejs.j2c.ConnectionManager.allocateConnection(ConnectionManager.java:599)
at com.ibm.ws.rsadapter.jdbc.WSJdbcDataSource.getConnection(WSJdbcDataSource.java:439)
at com.ibm.ws.rsadapter.jdbc.WSJdbcDataSource.getConnection(WSJdbcDataSource.java:408)
Any tips on this? I suspect I'm using a wrong class from hsqldb, or maybe my JDBC url is wrong...
In the example given in BDCP docs, the org.hsqldb.jdbcDriver class is used as the driver. The org.hsqldb.jdbc.JDBCDriver is supported only in HSQLDB 2.x, but the other class is supported by all versions of HSQLDB.