JDBCStorageService

Overview

The JDBCStorageService is a database-level compatible replacement for the JPAStorageService and provides a Storage Service on top of an RDBMS. It communicates directly with the database rather than using Hibernate ORM which has issues with reliability and the lack of provenance of its software artifacts.

It is possible to swap between the JPA and JDBC storage service, and indeed to have different versions running on different nodes.

The JPA Storage Service will be removed in V5 of the IdP, due in 2023. Please migrate prior to that point.

Plugin Installation

Starting with IdP 4.2 you can the install the latest plugin version supported on your IdP version with
.\plugin.sh -I net.shibboleth.plugin.storage.jdbc

Plugin ID	Module(s)	Latest Version	Bug Reporting
net.shibboleth.plugin.storage.jdbc	None	1.0.4; download	https://shibboleth.atlassian.net/browse/JJDBC

For a detailed guide on how to install plugins, see here.

In summary, use the plugin command that ships with the IdP to install the plugin from either a local file pre-downloaded, from a URL or by pluginId^4.2

Installation

C:>\opt\shibboleth-idp\bin\plugin.bat -I net.shibboleth.plugin.storage.jdbc

or

$ /opt/shibboleth-idp/bin/plugin.sh -i http://shibboleth.net/downloads/identity-provider/plugins/pluginName/version/URL

Switching from JPAStorageService

If you are currently running with the JPAStorageService you can reconfigure to use the JDBCStorageService relatively easily

Locate the configuration (search for the class name org.opensaml.storage.impl.JPAStorageService
Remove the EntityManagerFactory bean, taking note of the dataSource property.
Remove also the VendorAdapter bean if present.
Change the StorageService bean
- Replace class="org.opensaml.storage.impl.JPAStorageService" with parent="shibboleth.JDBCStorageService"
- Remove the constructor parameter and instead add a pointer to the dataSource you noted above p:dataSource-ref="...."

At this stage you should be able to test the configuration.

Once it works you can change bean names appropriately and add any extra configuration as detailed below.

Database Preparation

If you are not upgrading from a JPAStorageService configuration you need to:

Create the database table for the plugin to use.
Download the appropriate JDBC driver.
(Optionally, but recommended) Download a Connection Pooling implementation.

If you are moving from the JPAStorageService you do not need to make any changes to your database and you can use the same configuration for the DataSource as you did for the JPAStorageService.

Creating the Database

Example Schemas are shown below

Whatever you do, you MUST ensure the context and id columns are case-sensitively handled and compared. That is a requirement of the API that will be using the database. This is frequently NOT the default behavior of databases such as MySQL.

The specific examples that follow should NOT be assumed to be functional, as they likely are the product of different sources, varying amounts of testing (including none), and may not be current. Drivers get updated frequently and JDBC and database bugs appear and disappear with regularity. When in doubt, always grab new ones when problems appear.

MySQL

CREATE TABLE storagerecords (
  context varchar(255) NOT NULL,
  id varchar(255) NOT NULL,
  expires bigint DEFAULT NULL,
  value text NOT NULL,
  version bigint NOT NULL,
  PRIMARY KEY (context, id)
);

PostgreSQL or H2

CREATE TABLE storagerecords (
  context varchar(255) NOT NULL,
  id varchar(255) NOT NULL,
  expires bigint DEFAULT NULL,
  value text NOT NULL,
  version bigint NOT NULL,
  PRIMARY KEY (context, id)
);

Oracle

CREATE TABLE storagerecords (
  context varchar2(255) NOT NULL,
  id varchar2(255) NOT NULL,
  expires number(19,0),
  value clob NOT NULL,
  version number(19,0) NOT NULL,
  PRIMARY KEY (context, id)
);

The value column must be arbitrarily wide to allow the JDBC storage service to back IdP Sessions

JDBC Driver

You need to locate, download and verify the JDBC driver for your database and place it in edit-webapp/WEB-INF/lib. After populating edit-webapp/WEB-INF/lib you should execute bin/build.sh or bin/build.bat as appropriate for your environment.

Connection Pooling

We recommend the use of a DataSource that provides connection pooling, which may require installing an additional library as well.

The following libraries provide connection pooling functionality:

Having located, downloaded and verified the connection pooling jar you should place it in edit-webapp/WEB-INF/lib. After populating edit-webapp/WEB-INF/lib you should execute bin/build.sh or bin/build.bat as appropriate for your environment

Configuration

You need to add the definition of a bean derived from shibboleth.JDBCStorageService into an appropriate configuration file (usually global.xml). The options you can provide to the bear are detailed below.

Behavioral Options

The behavior of the Storage Service is controlled by the following options

Option Property Name	Default	Description
dataSource	Required	The DataSource to use
cleanupInterval	“PT10M” (or the value of the property `idp.storage.cleanupInterval` if it is set)	The time between one cleanup and another. A value of 0 indicates that no cleanup will be performed.
retryableErrors		A comma separate list of SQL errors which will cause a failed transaction to be retried (a maximum of transactionRetry times)
transactionIsolation	8 (Connection.TRANSACTION_SERIALIZABLE)	The level of transactional isolation required as described for the Connection Interface Starting in V2.0.0, if 0 (TRANSACTION_NONE) is specified then the transactional isolation is not set at the Connection level
transactionRetries	3	Number of retries if insertion fails due to database transaction bugs
verify	true	Whether to verify the database connection on startup
localLocking	false	Whether to do thread level locking to arbitrate access (for this IdP) to the the database. This can be useful in high contention situations when multiple transaction retries are happening.
contextSize	255	The size of the ‘context’ column in you database. Only change this if you are using a non-standard DDI
keySize	255	The size of the ‘key’ column in you database. Only change this if you are using a non-standard DDI
valueSize	`Integer.MAX_SIZE` (2³¹)	The mazimum size of the ‘value’ column in you database. Only change this if you are using a non-standard DDI

SQL Options

It is possible to redefine any or all all the SQL statements that are sent to the database by the JDBCStorageService. The following is the list

Option Property Name	Default	Notes
preCreateQuerySQL	SELECT expires FROM StorageRecords WHERE context =? AND id=?	The SQL to query the state of the table prior to creating a new record. Issued in the same transaction as createCreateRecordSQL or createUpdateRecordSQL
createCreateRecordSQL	INSERT INTO StorageRecords(context, id, expires, value, version) VALUES (?, ?, ?, ?, 1)	The SQL to create a new record. Issued in the same transaction as preCreateQuerySQL
createUpdateRecordSQL	UPDATE StorageRecords SET value=?, version=1, expires=? WHERE context=? AND id=?	The SQL to create a update an expired record (instead of a create) Issued in the same transaction as preCreateQuerySQL
deleteByContextExpiredSQL	DELETE FROM StorageRecords WHERE context = ? AND expires < ?	The SQL to “reap” away expired records for a given context
deleteByContextSQL	DELETE FROM StorageRecords WHERE context = ?	The SQL to remove all records for a given context
deleteByExpiredSQL	DELETE FROM StorageRecords WHERE expires < ?	The SQL to remove all expired records (as part of the cleanup task)
preDeleteQuerySQL	SELECT version FROM StorageRecords WHERE context =? AND id=?	The SQL to determine whether the a record is the correct one to be deleted. Issues in the same transaction as deleteRecordSQL
deleteRecordSQL	DELETE FROM StorageRecords WHERE context=? AND id=?	The SQL to delete a specific record. Issued in the same transaction as preDeleteQuerySQL
preUpdateQuerySQL	SELECT version, expires, value FROM StorageRecords WHERE context =? AND id=?	The SQL to determine the state of a record prior to its update Issued in the same transaction as updateRecordSQL
updateRecordSQL	UPDATE StorageRecords SET value=?, version=?, expires=? WHERE context=? AND id=?	The SQL to update a specific record Issued in the same transaction as preUpdateQuerySQL
readAllByContextSQL	SELECT id, expires, value, version FROM StorageRecords WHERE context = ?	The SQL to return all the records associated with a specific context
readAllSQL	SELECT context, id, expires, value, version FROM StorageRecords	The SQL to return all the records
readContextsSQL	SELECT context FROM StorageRecords	The SQL to return all the context names
readRecordSQL	SELECT version, expires, value FROM StorageRecords WHERE context =? AND id=?	The SQL to read a specified record.
updateExpiresByContextSQL	UPDATE StorageRecords SET expires = ? WHERE context = ? AND expires > ?	The SQL to refresh the expiration of all currently unexpired records.

Examples

In the example below use of Commons DBCP is demonstrated (class="org.apache.commons.dbcp.BasicDataSource", p:url="..." in the DataSource bean). When using other Connection Pool implementations change the class and properties appropriately, e.g.:

Tomcat DBCP2: class="org.apache.tomcat.dbcp.dbcp2.BasicDataSource", p:url="..."
Tomcat JDBC Pool: class="org.apache.tomcat.jdbc.pool.DataSource", p:url="..."
HikariCP: class="com.zaxxer.hikari.HikariDataSource", p:jdbcUrl="..."

    <bean id="my.dataSource" class="org.apache.commons.dbcp2.BasicDataSource" destroy-method="close" lazy-init="true"
       p:driverClassName="......"
       p:url="jdbc:hsqldb:mem:StorageService"
       p:username="shibboleth"
       p:retryableErrors="4001, 4002"
       p:password="%{JDBCPassword}" />

    <bean id="JDBCStorageService" parent="shibboleth.JDBCStorageService"
          p:dataSource-ref="dataSource"
          p:transactionIsolation="4"
          p:retryableErrors="40001"
     />

Example conversion from JPAStorageService

Old JPA Configuration

<bean id="shibboleth.JPAStorageService"
        class="org.opensaml.storage.impl.JPAStorageService"
        p:cleanupInterval="%{idp.storage.cleanupInterval:PT10M}"
        c:factory-ref="shibboleth.JPAStorageService.EntityManagerFactory" />

<bean id="shibboleth.JPAStorageService.EntityManagerFactory"
    class="org.springframework.orm.jpa.LocalContainerEntityManagerFactoryBean">
    <property name="persistenceUnitName" value="storageservice" />
    <property name="packagesToScan" value="org.opensaml.storage.impl" />
    <property name="dataSource" ref="shibboleth.JPAStorageService.DataSource" />
    <property name="jpaVendorAdapter" ref="shibboleth.JPAStorageService.JPAVendorAdapter" />
    <property name="jpaDialect">
        <bean class="org.springframework.orm.jpa.vendor.HibernateJpaDialect" />
    </property>
</bean>

<bean id="shibboleth.JPAStorageService.JPAVendorAdapter"
    class="org.springframework.orm.jpa.vendor.HibernateJpaVendorAdapter">
    <property name="database" value="MYSQL" />
</bean>

After Initial conversion

<bean id="shibboleth.JPAStorageService" 
      parent="shibboleth.JDBCStorageService"
      p:cleanupInterval="%{idp.storage.cleanupInterval:PT10M}"
      p:dataSource-ref="shibboleth.JPAStorageService.DataSource"/>