Blocker Description
Currently we have a bug where, if a single batch/operation has two updates for a single row, and that row has already been flushed, then we try to make two inserts into a DMS with the same key at the same timestamp, resulting in an assertion failure:
^[[0;31mF0910 17:37:15.487797 18644 deltamemstore.cc:74] Check failed: !mutation.exists() Already have an entry for rowid 23 at timestamp 9
