Overview
After job ODQ_DAEMON_CLIENT_nnn (program ODQ_DAEMON) is started, it runs for 15 minutes and checks every 15 seconds (Takt frequency).
Every 15 minutes a new job is started after finishing of the first.
Every 15 seconds, it queries the real time datasources to see if there is new data.
2 main tasks of ODQ_DAEMON
a.) Extracting the data to the operational delta queue
b.) Sending notifications to BW that new data is in the queues
Whether the first step a.) is performed depends on how the DataSource delta works:
- For DataSources which push the delta directly into queue (typical in Logistics or SLT), the extraction is not needed.
- For DataSources with Pull-delta and extraction job, the extraction part is used to write the data to the queue.
So step a.) depends on the involved DataSources and is optional.
In the notification step b.), RFC-calls are done to notify BW and to trigger/inform there the real-time processing.
Job log details
Sample:
17.03.2021 00:12:18 Notification task 1 started on server ldcidwt_DWT_00
17.03.2021 00:12:19 Extractions task 1 started on server ldcidwt_DWT_00
17.03.2021 00:12:19 Queue ABAP_CDS\RSDAEMONCDS$F will be assigned to extraction task 1
17.03.2021 00:12:30 Takt 1 to TSN {2021-03-17 00:12:30 000007 CET}. Utilization: extraction 41%, notification 9%
The percentage number of extraction and notification in the job log relates to the task time set in the demon (Default is 15s). This means if the notifications triggered by the demon job would run in total 7,5s within these 15 seconds, you would see the percentage of 50% in the job log. Same logic for the extraction task.
In general streaming for DataSources with extraction logic makes only sense if the delta extraction is fast. If one delta extractor takes e.g. some minutes, it does not make sense to use the realtime demon which tries to trigger the delta every 15 seconds. It cannot trigger the delta, as the last extraction is still running. In this case you will see 100% or even more than 100% in the job log.