19.3    Housekeeping

The standard housekeeping activities can be executed by using the background processing within the AEX. Background jobs can be accessed via SAP NetWeaver Administrator (navigate to SOA • Monitoring • Background Job Processing Monitor).

The next section will explore each of the background jobs.

19.3.1    Archiving

Message archiving provides the ability to archive messages that have been processed successfully or manually edited to an archiving space.

There are two archive sessions required to archive a message:

• The first session writes the message to the archive storage.
• The second session removes the archived message from the persistence layer, which is the database of the AEX.

19.3.2    Deletion

Within SAP PO, only messages that have fulfilled the following criteria will automatically be deleted from the database by the deletion job:

• Messages that have successfully been processed
• Messages that have reached their retention period
• Messages that aren’t candidates for archiving

19.3.3    Restarting

The restarting job can be scheduled to automatically restart messages that are in an error status. This is applicable for all messages for which the number of defined restart attempts has been exceeded. The restart attempts attribute specifies how many times a failing message can be retried before abandoning it. For restarting messages, the following parameters can be used:

• MAX-RESTART
  This parameter defines the maximum number of restart attempts for messages in status error or system error.
• RESTART-BLACKLISTED
  This parameter is used for jobs that need to reprocess messages that have been blacklisted and have a nondeliverable (NDLV) state status.

Blacklisting

Blacklisting is a mechanism that can be used to prevent situations in which individual messages can cause the server to overload and even fail.

For example, a message with a big payload (in terms of size) can cause the server to crash. Blacklisting is the process of identifying and isolating messages that can potentially harm the system or messages that were being processed when there was a server failure.

If a Java server node fails, and a message can’t be processed successfully after two attempts, then the messages will be assigned an NDLV status. When messages are in an NDLV state, an administrator must decide whether to allow the message to be reprocessed or to put the message in a FAIL state.

19.3.4    Recovery

This type of job is used to recover jobs that are lost as a result of a database failure. By default, the recovery job runs once per day, but additional recovery jobs can be scheduled if required.