Are there any tuning points to improve Qlik Replicate performance when using Oracle XStream as a source endpoint?
When using Oracle as the source endpoint, the apply throughput is 46,000–48,000 records per second. With Oracle XStream as the source endpoint, it is 7,000–9,000 records per second.
Are there any Qlik Replicate configuration settings (e.g., internal parameters) to improve processing performance?
This is a fascinating topic! Let me clarify the scenario to ensure I understand correctly: You performed 1,000,000 updates in an Oracle database and then compared the performance of two endpoints by initiating Change Processing tasks from the same specific timestamp on both. Is that correct? Could you elaborate on the exact steps in your tests? I'd like to understand them better and attempt to reproduce the scenario in my local environment.
Thanks,
John.
Help users find answers! Do not forget to mark a solution that worked for you! If already marked, give it a thumbs up!
Two tables with about 20,000,000 records each were updated with one-by-one commits on the source DB server.
1. source endpoint is Oracle 2. source endpoint is Oracle XStream
After completing the first test, the second test was performed. The result showed that the performance of the second test (Oracle XStream) was lower. In our internal test environment, Oracle XStream performed better, but in the POC environment, its performance was lower.
The POC environment DB has an SGA size of 50GB and STREAMS_POOL_SIZE of 10GB. Could the lower performance be related to the memory size?
Yes Source Oracle Parameter Does play in such scenarios as Qlik replicate is an application in which Submitts Querry to Db and DB has to process the request and revert it to QR.
you can make changes to STREAMS_POOL_SIZE to Same as POC env and then check and if you have SGA is ower then for an update it will take longer in compare to higher SGA values.
The internal test environment has smaller SGA memory and STREAMS_POOL_SIZE values, where the Oracle XStream endpoint performs faster than the Oracle endpoint.
In the POC environment with larger SGA memory and STREAMS_POOL_SIZE values, the Oracle XStream endpoint is slower than the Oracle endpoint.
The delay occurs in the source database.
Are there ways to improve performance when using Oracle XStream? (For example, increasing internal buffer values, etc.)
