Hi, I read a description of MVCC in http://www.cs.ust.hk/~dimitris/CS530/L24.ppt and found that this isn't exactly what is implemented in PostgreSQL. For example, for a sequence of operations like: 1: T1 sets isolation to serializable & begins a transaction 2: T2 sets isolation to serializable & begins a transaction 3: T1 reads X into v1 4: T2 reads Y into v2 5: T1 writes v1 into Y 6: T2 writes v2 into X 7: T1 commits 8: T2 commits Obviously, this sequence is also not a serializable execution. However, it is allowed by PostgreSQL. Moreover, according to the MVCC reference above, step 5 should really fail because the read timestamp of Y is that of T2, which is greater than that of T1. I understand that PostgreSQL doesn't implement predictive locking, but this example doesn't involve any phantom at all. It is plain multiversion timestamp concurency control. Any idea? Thanks in advance! ----- -- Kent Tong Wicket tutorials freely available at http://www.agileskills2.org/EWDW Axis2 tutorials freely available at http://www.agileskills2.org/DWSAA -- View this message in context: http://www.nabble.com/postgresql%27s-MVCC-implementation-tp18302020p18302020.html Sent from the PostgreSQL - general mailing list archive at Nabble.com.
