[Pgcluster-general] technical question

[Andreas Kostyrka]

Hi!

As pgcluster uses statement replications, I've wondered how it deals
with the usual problems of MVCC-based systems in relation to statement
replication?

E.g. How does it deal with multiple READ-COMMITED transaction? These
are strictly timing sensitive?

INSERT INTO history (_time, id, value) VALUES (now(), 10, (SELECT somevalue
FROM table WHERE id=10));

Andreas