Background documentation Starting and Stopping 

The following table provides you with an overview of where and how to start and stop the various tools and engines required for SAP NetWeaver PI.

Tools for Starting and Stopping

Tool/Engine

Access

Detailed Description

System Landscape Directory (SLD)

Choose Administration Start/Stop SLD at http://<host>:<port>/sld

After you have installed SAP NetWeaver PI, you must configure the SLD and then start it manually. Afterwards it restarts automatically if it stops at any point.

You can also start the SLD explicitly, for example, when AS Java is restarted or when there is a request for the SLD.

After you have started the SLD, you must ensure that the self registration tool for the technical systems is running.

Integration Server (including Business Process Engine and Integration Engine)

No special tool; started/stopped automatically when its host starts/stops.

When the Integration Server is started, it registers itself in the SLD.

Enterprise Services Builder and Integration Builder

No special tool; started/stopped automatically when their host starts/stops.

When the Enterprise Services Builder or the Integration Builder is started, it registers itself in the SLD.

Runtime Workbench

No special tool; started/stopped automatically when its host starts/stops.

When the Runtime Workbench is started, it registers itself in the SLD.

Adapter Engine

No special tool; started/stopped automatically when its host starts/stops.

When the Adapter Engine is started, it registers itself in the SLD.

Note

Basic configuration steps are performed during and after installation. The SAP NetWeaver Installation Guide and the basic configuration of PI explain these steps and describe checks to verify that the landscape is functioning properly.

Stop Procedure

To stop the various components of your PI system, perform the following steps:

...

  1.  If you use the Business Process Engine (BPE), first check how to start and stop the BPE.

More information: Starting and Stopping the Business Process Engine.

  2.  Stop the message flow from your sending systems where possible. If the sending system uses proxies, you can follow steps 5 to 8 to stop the message flow. In all other cases continue with step 3.

  3.  Stop all sender communication channels on your central Advanced Adapter Engine to avoid new messages coming in. Otherwise, messages might still be picked up and lost if a point-in-time recovery of your database is required, for example, for patching, upgrade, or hardware switch.

Proceed as follows:

...

  a.  Start the Runtime Workbench at http://<host>:<port>/rwb.

  b.  Choose Component Monitoring and display all components.

  c.  Choose your central Adapter Engine and then choose Communication Channel Monitoring.

  d.  In the filter, select the direction Sender and use the filter.

  e.  In the list of selected communication channels, choose Multiple Selection On and select all listed channels.

  f.  Choose Stop.

  4.  Monitor the status of your central Adapter Engine until all incoming messages are processed successfully. Otherwise these messages might fail and you would have to restart them as described in step 9 of the start procedure below.

Proceed as follows:

...

  a.  Start the Runtime Workbench at http://<host>:<port>/rwb.

  b.  Choose Component Monitoring and display all components.

  c.  Choose your central Adapter Engine and then choose Engine Status.

  d.  Refresh the Backlog tab page until the total backlog of the sender services in 0.

  e.  Continue with step 9.

  5.  Lock the Integration Server for incoming messages by calling the transaction Integration Engine – Administration (SXMB_ADM) and choosing Integration Engine Configuration Specific Configuration Change New Entries. Select the category RUNTIME and then the parameter ENTRY LOCK: Set the current value to 1 (LOCKED) and choose Save.

Note

If you set this parameter, messages that have already been received will still be processed. You have to wait until there are no more PI outbound queues in transaction SMQ2 and until all messages are processed by the central Adapter Engine.

  6.  Deregister the queues of your PI system by calling the transaction Integration Engine – Administration (SXMB_ADM) and choosing Manage Queues. Select Deregister Queues and choose Execute action. This immediately stops the processing of all messages in these queues

  7.  Monitor the remaining entries in the queues by calling the qRFC Monitor Inbound Queue (transaction SMQ2) until no more queues are listed.

  8.  Check the tRFC entries by calling the transaction Transactional RFC (SM58) until the list is empty.

  9.  Decide if you want to stop the non-central Adapter Engine (if used at all). This is an optional step with the following advantages/disadvantages:

  Advantage of stopping

No errors occur during the downtime of the Integration Server. However, the sending systems are not able to address the Adapter Engine and no files are polled from the file system. This is especially important if a point-in-time recovery of your database is required, for example, for patching, upgrade, or hardware switch.

Proceed as described in step 3.

  Advantage of not stopping

All messages with quality of service EO (Exactly Once) and EOIO (Exactly Once in Order) that enter the non-central Adapter Engine are persisted during the downtime of the Integration Server. However, they have to be manually restarted afterwards.

  10.  Stop the Integration Server.

Start Procedure

To start the various components of your PI system, perform the following steps:

...

  1.  If the SLD is running on a different host and has been stopped for maintenance reasons, you need to start it.

  2.  Start the Integration Server.

  3.  Wait until AS Java and all PI-related services are started.

This process can be monitored with the SAP NetWeaver Administrator.

Choose Operations Management Infrastructure Management Start & Stop Java EE Applications.

  4.  Restart the non-central Adapter Engine (if used and stopped at all). If you did not stop it, access the message monitor and manually restart the messages with errors.

  5.  Register the PI queues by calling the transaction Integration Engine – Administration (SXMB_ADM) and choosing Manage Queues. Select Register Queues and choose Execute action.

  6.  If you have locked the Integration Server for incoming messages, open it by calling the transaction Integration Engine – Administration (SXMB_ADM) and choosing Integration Engine Configuration Specific Configuration Change New Entries. Select the category RUNTIME and then the parameter ENTRY LOCK: Set the current value to 0 (NOT LOCKED) and choose Save.

  7.  Call the qRFC Monitor Inbound Queue (transaction SMQ2) and check the status of the queues. The correct status is "RUNNING". It might take several minutes before this status appears.

  8.  If you have stopped the sender communication channels on your central Adapter Engine, start them to unlock message processing on the Adapter Engine.

Proceed as follows:

...

  a.  Start the Runtime Workbench at http://<host>:<port>/rwb.

  b.  Choose Component Monitoring and display all components.

  c.  Choose your central Adapter Engine and then choose Communication Channel Monitoring.

  d.  In the filter, select the direction Sender and use the filter.

  e.  In the list of selected communication channels, choose Multiple Selection On and select all listed channels.

Note

If for whatever reason some of the listed channels were stopped before you stopped you PI system, you must not select them for restart now.

  f.  Choose Start.

  g.  Check whether all communication channels have a green status after their restart.

  9.  Check for messages with an error status in the central Integration Engine, in the Adapter Engines, and in your business systems (if used) as follows:

  a.  Start the Runtime Workbench at http://<host>:<port>/rwb.

  b.  Choose Message Monitoring.

  c.  Select the following options as required:

  Integration Engine <SID>

  Adapter Engine <SID>

  Proxy Runtime <SID_client>

  d.  Select Database (Overview) and choose Display.

Here you can easily identify messages with errors and manually repeat or cancel them where possible. For messages processed on an Adapter Engine, you can also use an automatic restart job. Proceed as follows:

...

  a.  Start the Runtime Workbench at http://<host>:<port>/rwb.

  b.  Choose Component Monitoring and display all components.

  c.  Choose the Adapter Engine in question and then choose Background Processing.

  d.  Use an existing Restart job or create a new one if required.

  10.  Check your integration processes (if used) by calling the transaction Business Process Engine – Monitoring (SXMB_MONI_BPE) and choosing Restart Process After System Crash.

  11.  If a cache refresh was performed and was still running when the Integration Server was shut down, call the transaction Integration Directory Cache (SXI_CACHE) and check for errors by choosing Runtime Cache Display Refresh Error.

  12.  If your sending systems were stopped, you need to restart them; if they use ABAP proxies, apply steps 5 to 7.

  13.  Check for errors on all sending systems.

  14.  If you use the BPE, check also how to start and stop the BPE.

More information: Starting and Stopping the Business Process Engine.

 

Was this page helpful to you?