Re: Anyone working on better transaction locking?

"Andrew Sullivan" <andrew(at)libertyrms(dot)info> wrote in message
news:20030408230518(dot)GB32207(at)libertyrms(dot)info(dot)(dot)(dot)
> On Tue, Apr 08, 2003 at 01:45:25PM +0000, Ron Peacetree wrote:
> > Unfortunately, the performance of PostgreSQL MVCC in comparison to
> > say Oracle (the performance leader amongst MVCC DB's, and pretty
much
> > for all DB's for that matter) is not competitive. Therefore there
is
>
> What, is this a troll?
Time will tell. Hopefully not.

> The question apparently reduces to, "Why isn't PostgreSQL
> as good as Oracle?"
Actually, you've just used reductio absurdium, not I. My question
compares PostgreSQL to the performance leaders within this domain
since I'll have to justify my decisions to my bosses based on such
comparisons. If you think that is unrealistic, then I wish I worked
where you do. If you think that is unreasonable, then I think you're
treating PostgreSQL as a religion and not a SW product that must
compete against every other DB solution in the real world in order to
be relevant or even survive.

> 1. For what? There are things that Oracle users will tell you
> not to do, because there is a faster way in Oracle.
>
> 2. How do you know? I haven't seen any real benchmarks
> comparing PostgreSQL and Oracle similarly tuned on similar hardware.
> So I'm sceptical.
Please see my response(s) to Tom below.

> But if you have specifica areas which you think need improvement
> (and aren't already listed in the TODO), I'll bet people would like
to
> hear about it.
Please see my posts with regards to sorting and searching, two phase
execution, and two phase commit. I'll mention thread support in
passing, and I'll be bringing up other stuff as I investigate. Then
I'll hopefully start helping to solve some of the outstanding issues
in priority order...

"Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote in message
news:4096(dot)1049860699(at)sss(dot)pgh(dot)pa(dot)us(dot)(dot)(dot)
> Ron, the tests that I've seen offer no support for that thesis.
What tests? I've seen no tests doing head-to-head,
feature-for-feature comparisons (particularly for low level features
like locking) of PostgreSQL vs the "biggies": DB2, Oracle, and SQL
Server. What data I have been able to find is application level, and
certainly not head-to-head. From those performance results, I've had
to try and extrapolate likely causes from behavioral characteristics,
docs, and what internal code I can look at (clearly not much from the
"biggies").

If you have specific head-to-head, feature-for-feature comparison test
results to share, PLEASE do so. I need the data.

> If you want us to accept such a blanket statement as fact, you'd
> better back it up with evidence. Let's see some test cases.
Soon as I have the HW and SW to do so, it'll happen. I have some "bet
the company" decisions to make in the DB realm.

Test cases are, of course, not the only possible evidence. I'll get
back to you and the list on this.

> Postgres certainly has plenty of performance issues, but I have no
> reason to believe that the fundamental MVCC mechanism is one of
> them.
Where in your opinion are they then? How bad are they in comparison
to MySQL or any of the "Big Three"?