Has anyone seen any speed issues in NWMKDE with an NSS SYS volume? Before anyone asks - all the BM caching and logging is to a traditional volume

I'm seeing slow access through one of our proxies with high utilisation on proc 0 mainly tied up by NWMKDE. Breaking down NWMKDE's stack shows it waiting for NSS to respond from CacheFlushMyCacheBufs.

I've tried shutting off indexed logging which has improved performance but during high usage is still causing slow access.