Gordon Sim wrote: > In the case of a single node, if that node is killed in the middle of > processing a tx-commit and then restarted, are there any atomicity > guarantees? Could some of the operations be completed with others being > lost? The latter. Matthias.