Currently the oplogTruncateAfterPoint is getting redundantly written to even when the value is not changing – imagine a quiet system, still periodically running the code to flush data to disk (nothing to flush, of course), before which we always do an oplogTruncateAfterPoint update regardless of anything.