I am loading multiple batches of data (~40 MB) in parallel to the Oracle DB. I am noticing parallel Oracle commits to DB are very slow in that case. But single commit is much faster. Looks like this problem is related to the Oracle redo log files or log file syncs. Could you help to solve this problem.
If you have log file switch (checkpoint incomplete) events, it's means that your sizing your redo-logs to small. What is the size of them? Could you add more with biggest size(it's easy to do it)?