Authentication Session Persistence

This topic is regarding sessions in the WSO2 Identity Server (IS) and the process of enabling session persistence for these sessions. This is particularly useful when the remember me option is selected when logging into either the service provider or the WSO2 Identity Server.

Understanding sessions in the WSO2 Identity Server

When you log in to the Web application using WSO2 Identity Server, a single sign-on (SSO) session is created by the Identity Server for the user of the application. Also, as the user logs in to the Web application, there is session created in the Web application itself for the user. These are two separate sessions and they are not synchronized with each other.

For an example, if the Web application has a session timeout of 20 minutes and the WSO2 IS SSO session timeout is 5 minutes, the application users will not see the login page up to 20 minutes. That is because WSO2 IS is not invalidating the session by force on the Web application.

If the Web application session timeout is 5 minutes and the WSO2 IS SSO session timeout is 20 minutes, the users will not see the login page up to 20 minutes. This is because even when the Web application container session timeout after 5 minutes, the session was kept alive since WSO2 IS SSO session is still alive. The user will not be redirected to the login screen of the SP until the WSO2 IS SSO session is invalidated.

WSO2 Identity Server creates separate SSO session for SSO login and it is different from the session that is created when you log in to the Identity Server management console.

When end user logs in through the WSO2 Identity Server for the service provider application (using SAML2 SSO, OpenID Connect, Passive STS, etc.), the Identity Server creates a SSO session for end users and a cookie that is related to the created SSO session is set to the user’s browser.

This cookie can be seen as commonauthId. It is set to the user’s browser with the hostname of WSO2 Identity Server instance and the value of the commonauthId cookie is the SSO session identifier. When SSO session is created in the WSO2 Identity Server, the session is put into the session cache and persisted to the database. To persist it in to the database, you must enable the session persistence.

Click here to read about the importance of persisting the session.

SSO sessions have been stored in an in-memory cache. It is recommended to persist the SSO session due to following reasons.

If you are running a single WSO2 Identity Server instance and the server is restarted, all SSO session would be removed. If you have multiple nodes of WSO2 instances, It is not guaranteed that you can recover all the sessions. Although the cache is distributed, it is not 100% split to each node.
Cache has a limit. If there are large number of SSO sessions, memory can be high and server performance may reduce. So usually the cache is evicted after a given number of entries (by default 10000 entries). Therefore, some SSO session can be evicted from caches when there are large number of user logins.
When there is a clustered development, if you have no persistence, you need to rely completely on the distributed cache. However, if you have persistence, you can rely on it as well. This increases the reliability of the overall system.

The following configuration found in the <IS_HOME>/repository/conf/identity/identity.xml file, under the the Server and JDBCPersistenceManager elements is used to enable session persistence.

<SessionDataPersist>
	<Enable>true</Enable>
	<Temporary>true</Temporary>
	<PoolSize>0</PoolSize>
	<SessionDataCleanUp>
		<Enable>true</Enable>
		<CleanUpTimeout>20160</CleanUpTimeout>
		<CleanUpPeriod>1140</CleanUpPeriod>
		<!--Instead of deleting all the records at once, we are deleting the records in chunks to prevent
		the -->
		<!--possible deadlock and lock scenarios. The following property defines the chunk size.-->
		<DeleteChunkSize>50000</DeleteChunkSize>
	</SessionDataCleanUp>
	<OperationDataCleanUp>
		<Enable>true</Enable>
	</OperationDataCleanUp>
</SessionDataPersist>

The following table describes the elements of the configurations mentioned above.

Configuration element	Description
Enable	This enables the persistence of session data. Therefore, this must be configured to `true` if you wish to enable session persistence.
Temporary	Setting this to `true` enables persistence of temporary caches that are created within an authentication request.
PoolSize	To improve performance, OAuth2 access tokens are persisted asynchronously in the database using a thread pool. This value refers to the number of threads in that thread pool.
SessionDataCleanUp	This section of the configuration is related to the cleaning up of session data.
Enable	Selecting true here enables the cleanup task and ensures that it starts running.
CleanUpTimeOut	This is the timeout value (in minutes) of the session data that is removed by the cleanup task. The default value is 2 weeks.
CleanUpPeriod	This is the time period (in minutes) that the cleanup task would run. The default value is 1 day. This is used for both session data cleanup and operation data cleanup through the same task.
DeleteChunkSize	This value determines (limits) the number of rows that will be deleted in a single delete query statement. The default value is 50000. If the number of rows to delete is larger than this limit, the DELETE statement is repeated until the number of affected rows is less than the LIMIT value.
OperationDataCleanUp	This section of the configuration is related to the cleaning up of operation data.

About size of Delete Chunks

The value of delete chunk size depends on various factors such as:
- The size of a row of the table. You can calculate this by adding bytes needed for each column, and some control information per record.
- The underlying physical storage. e.g Normal SATA spinning Disk, SSD, SCSI, Amazon S3/EBS etc.
- The type of the database engine (e.g. MySql, MSSql, Oracle). Each one has different data organization mechanism in physical storage.
- OS where the DB server runs.

The best way to calculate the chunk size it is to run a sample program that calculates the size. To do this, you may need to insert a large number of artificial records on the table and run a deletion query with incremental chunk sizes. Then you can plot the result in a table to see which is the optimal chunk size for a specific configuration.

Also, this chunk size configuration may change later if you decide to upgrade the DB cluster, their storage etc. Hence it is not a static value. The optimal value needs to be measured each time you do a change in your database infrastructure.

Note: Usually, the optimum chunk size is the power of 2 and typically multiple of 512. This can be used for incremental values.

Why do we need to have a proper delete chunk size?

In the world of World Wide Web, Sessions are the simplest way to store data for individual users against a unique session ID. These can be used to persist state information between page requests. When we consider all the requests and responses that comes to a page per day, that is really a large amount. Due to this reason, the session data storing in the “IDN_AUTH_SESSION_STORE” table in the WSO2CARBON_DB of WSO2 IS is really high. This table fills up quickly if your system receives too many loads of requests.

After a certain period of time these session data is not really necessary. In order to stop the exponential table growth, we run a Session Clean Up Tasks in predefined time intervals via a script. If this storing session data is huge, the data that need to be deleted also will come in bulk.

The session clean up task also takes a certain amount of time and consumes certain amount of resources. Therefore, we always need need to improve the performance of session clean up tasks. For this DeleteChunkSize property has introduced in WSO2 Identity Server. This will also help to reduce dead-locks happening during session clean up task. With the use of DeleteChunkSize value, the deletion happens batch wisely. “DeleteChunkSize” of data is deleted at one time.

But if you didn’t select a correct number for the DeleteChunkSize, you will end up with less performance. Therefore selecting suitable chunk size for you database is a must.

This chuck size may differ due to various factors of your environment

The size of a row of the table. Calculate this by adding bytes needed for each column, and some control information per record.
The underlying physical storage. e.g Normal SATA spinning Disk, SSD, SCSI, Amazon S3/EBS etc.
The type of the database engine (e.g. MySql, MSSql, Oracle). Each one has different data organization mechanism in physical storage.
OS where the DB server runs.

Therefore, it is hard to select a DeleteChunkSize without considering all the above factors.