How are other retailers configuring Oracle Retail on RAC?

During volume testing of the batch we’ve experienced performance issues on certain jobs due to block and/or resource contention across the interconnect. Specifically we’re seeing high wait times related to the following events - gc buffer busy acquire, gc current block busy, gc cr failure.  We’ve tried following some standard RAC tuning recommendations like increasing the cache size on ‘hot’ sequences but it hasn’t resolved the problem(s).  When we pinned the batch to one node these wait events became negligible. The downside is I can’t leverage the resources on the 2nd node when I do this. We’re running v14.0.2 on