Scheduling Indexing

To enable Enterprise Search to deliver search results from the connected system, the data must be indexed.

So that users can always search the latest data during production operations, you schedule regular indexing runs and trigger extraordinary indexing runs on demand.

Once the search object connector is created, the system prepares the first indexing run by creating process chains and the required indexes in TREX. During this process, the search object connector is displayed with the status Preparing in the Administration Cockpit.

Once preparation is completed, the search object connector assumes the status Prepared, and you can start or schedule indexing.

Scheduling Regular Indexing Runs

You can schedule regular indexing runs for the search object connectors at the object type level.

The system performs delta indexing for object types that support this, provided that you have not chosen full-indexing mode. During delta indexing, the system indexes only those objects that have been created, changed, or deleted since the last indexing run. This type of indexing creates less load on the system and does not impact routine searching.

Caution Caution

However, during the first delta indexing run, SAP Netweaver Enterprise Search shuts down the source system once only to ensure the consistency of the necessary extraction.

End of the caution.

In the case of other object types or if you choose a full-indexing mode, the systems performs full indexing. Full indexing creates a higher load on the system than delta indexing. Therefore, you should schedule the time for full indexing of large data volumes carefully. By selecting the full-indexing mode you can decide whether or not the index should remain searchable during the indexing process.

Recommendation

Schedule the intervals for regular indexing runs at the object type level. The interval depends on how often changes the objects in question are subject to change. For example, reference data such as material types require indexing less often than the actual materials.

End of the recommendation.

Performing Full Indexing On Demand

In the following cases, it can be necessary to perform an extraordinary full-indexing run in addition to regular indexing:

If you makes changes to the connector model that necessitate indexing the data again.
More information: Managing Search Object Connectors
If an index contains errors.

Procedure

Log on to the administration cockpit.
Choose the search object connectors that you want to schedule indexing for and then choose Actions Schedule Indexing .
Select one or more object types.

Note
The entry screen allows you to make settings across connectors for more than one object type at the same time.

End of the note.
Specify a Start Time.
If you want to perform indexing regularly, enter a Recurrence Period.

Example
If you specify tomorrow's date at 01:00:00 and one day as the recurrence period, the system performs the indexing run from the next night onwards, every subsequent night at one o'clock.

End of the example.
If you want to perform full indexing, choose an Indexing Mode.
If you do not choose a mode, the system performs delta indexing if possible; otherwise it performs full indexing.

When you select a mode, you decide whether the existing index is retained during full indexing and updated row by row or completely emptied and then refilled.
Choose OK to activate the indexing schedule.

Result

The search object connector is now scheduled for indexing and goes through the following statuses during the first indexing run:

Note Note

All indexing runs are performed in the background and are not always run immediately; therefore, it can take some time until the required action is run in the background. In the administration cockpit, you can choose Refresh to display the progress of the action using the status changes.

End of the note.

Status	Comments
`Scheduled for Indexing`	This is the status of the search object connector once you have scheduled it for indexing as described above.
`Extraction`	The system schedules a job for each node in the search object connector. These jobs start the process chains that were created in the preparation phase and wait until the process chains are completed. The process chains extract the data from the back-end system and write them into a PSA table on the Enterprise Search appliance until all the relevant data is extracted. During this phase, the search object connector has the status `Extracting`.
`Being Indexed`	Once all the data is in the table, the system reads it from there and passes it on to indexing. During indexing, the search object connector has the status `Indexing`.
`Active`	Once indexing is finished, the search object connector has the status `Active`
`Error while Extracting` or `Error while Indexing`	If an error occurs during extraction or indexing , the search object connector is displayed with the corresponding status.

Once the search object connector is active, users obtain search results in the Search UI.