-
Title
Capture performance in 6.0 -
Description
Capture performance is noticeably slower in Shareplex 6.0. The performance degradation is traced to capture parameter SP_OCT_OLOG_QUEUE_SIZE which in 6.0.0 is set to 100 (default). When this parameter is changed to 1000, it improves capture performance in 6.0 . For RAC, we need to adjust one more parameter called SP_OCT_OPS_QUEUE_SIZE which in 6.0 is set to 1000 (default). After increasing this parameter to 10000, it increased performance on RAC as well.
-
Resolution
Non-RAC:
Please set this parameter to 1000 and restart capture for this to take affect. Syntax below:
sp_ctrl>set paramSP_OCT_OLOG_QUEUE_SIZE 1000
RAC:
sp_ctrl> set param SP_OCT_OLOG_QUEUE_SIZE 5000
sp_ctrl> set param SP_OCT_OPS_QUEUE_SIZE 10000
Increasing these parameters may increase the memory usage but will speed up capture. You will need to keep in mind the number of nodes and the overall activity of the nodes in RAC. You may want to start with SP_OCT_OLOG_QUEUE_SIZE 2000 and gradually increase it to 5000 on RAC and monitor memory utilization on RAC configurations. -
Additional Information
There are some considerations which need to be kept in mind with this. First of all, the larger the value of SP_OCT_OLOG_QUEUE_SIZE and SP_OCT_OPS_QUEUE_SIZE, the more memory capture process will take, especially in RAC configurations, since SP_OCT_OLOG_QUEUE_SIZE is a thread-centric parameter.