On 2024-10-05 09:59:00 -0700, Adrian Klaver wrote: > On 10/5/24 02:14, Peter J. Holzer wrote: > > On 2024-09-25 18:09:44 -0400, Tom Lane wrote: > > > "Peter J. Holzer" <hjp-pg...@hjp.at> writes: > > > Admittedly, that would normally not be a very long interval if BEGIN > > > did both things ... but on a busy system you could lose the CPU for > > > awhile in between. > > > > Assuming that the system does have a global clock of sufficiently > > fine resolution which returns strictly monotonically increasing > > timestamps[1], I think the following is true: > > > > Every snapshot divides the set of transactions into two non-overlapping > > subsets: Those which have committed early enough that their effects are > > visible in the snapshot and those which haven't. Let's call the first set > > the "earlier" transactions and the second the "later" transactions. Let's > > call the current transaction c and any transaction in the earlier set e > > (we ignore the later transactions for now). > > > > Performing a commit and taking a snapshot take some time, but there > > should be a time t_C(e) in each commit and t_S(c) in the snapshot, such > > that t_C(e) < t_S(c) for each "earlier" transaction. > > Assuming t_C is time of commit and t_S is time of snapshot, is the > above not the crux of the matter? Namely when in the current > transaction the snapshot is actually taken. That would determine what > constitutes an earlier visible transaction relative to the current > transaction. In other words I am not seeing how this changes anything?
The important part is in the last paragraph:
> > If we choose the transaction_timestamp to be >= t_S, then
> > transaction_timestamp(e) < t_C(e) < t_S(c) <= transaction_timestamp(c)
> > and therefore
> > transaction_timestamp(e) < transaction_timestamp(c)
In PostgreSQL, transaction_timestamp is taken during BEGIN (as Greg
noted). If it was instead taken at the end of the snapshot, it would be
guaranteed to be later than any transaction_timestamp of an earlier
transaction.
Again, I'm not arguing for such a change, but I'm wondering if recording
transaction_timestamp just after the snapshot might be a safe change or
whether that might break some assumption that programmers can currently
make.
hp
--
_ | Peter J. Holzer | Story must make more sense than reality.
|_|_) | |
| | | h...@hjp.at | -- Charles Stross, "Creative writing
__/ | http://www.hjp.at/ | challenge!"
signature.asc
Description: PGP signature
