Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.databases.postgresql > #159

Re: Is PostgreSQL good?

Show all headers | View raw

On Tue, 28 Jun 2011 10:34:03 +0200, Laurenz Albe wrote:


>> I can only see that PostgreSQL is much less popular than MySQL and I
>> can see that it isn't getting better.
> 
> First, be patient. And second, popularity is not always the measure of
> quality (ever heard of Microsoft?).

Patient? Postgres came about in 1995. It's 2011 now. How much more 
patience is needed? As for Microsoft, it is very good at what it does. MS-
Word is a class in itself, head and shoulders above the competition. 
Microsoft still owns 95% or so of the desktop and office software market, 
despite repeated assaults by both Linux and Apple and the former 
opponents like OS/2, WordPerfect, Borland, Lotus and all other names that 
are now but waning memories. Other than that, I have very good 
experiences with Windows7. I wouldn't exactly take MS as an example for 
the lack of quality. Look at it this way: users are voting with their 
wallets. So far, MS is winning the contest by a wide margin, despite the 
existence of the free alternatives. That says something, doesn't it?

> Well, you said that you spent months learning PostgreSQL. That's an
> investment that can be measured in money, right?

Yes, it is. In this case, it was wasted money. I've wasted a year of my 
life. 

> 
> Commercial software has a different and enticing pricing model. You'll
> get the thing for a moderate price at first (or you get a crippled free
> version), then you realize that without (expensive) support you can't
> even get patches for the showstopping bugs, and the next time you'll
> have to negotiate license fees you'll have to pay more.

Depends on the vendor. There are many business models, ranging from Red 
Hat which allows you to download for free and pay for the support if you 
need it to Oracle.

> With open source software you may have a higher initial cost: you will
> have to spend more money on integrating it into your environment, you
> might have to invest more into education, you might have to develop
> add-ons that you need and that are not supplied. But once this is done,
> costs will drop rapidly, and in the long run you'll be cheaper.

Unless the software itself doesn't fulfill your needs. And that is what 
I've been trying to convey to Postgres community all along: hints are not 
optional. They are essential. That is why I have drawn the line in the 
sand.


> Furthermore, you'll get added values with open source: the ability to
> influence decisions (the only enhancement request I ever made for Oracle
> years ago did not even receive an answer), the ability to help yourself
> if you run into a bug or need a feature.

You mean the ability to help yourself if you need something like hints? 
What happens if the product is heavily influenced by bozos saying 
something "we don't need to do that because we're not for profit"?

> Maybe there are other reasons besides terrible performance? Do you have
> to pay for TPC? Is it expensive?

TPC is mainly paid for by the equipment manufacturers. Having the meanest 
database machine on the market allegedly sells boxes. Database vendors 
are usually not too thrilled about benchmarks because of the fear that in 
a few months the vendor will release a better benchmark by some other DB 
and leave them with a huge marketing problem. If Postgres was so fast, I 
assure you, someone would have selected it for a TPC benchmark. The fact 
that it didn't happen puts things in perspective.


> PostgreSQL has some of these things.
> Even if Oracle has some performance related features that PostgreSQL
> does not, that does not prove that it will always perform better. I'm
> sure that Oracle will sometimes be faster than PostgreSQL, but I don't
> accept a general statement like that.

Look no further than the Postgresql performance mailing list. At least 
once a month, there is a question why does Oracle do this or that faster 
than Postgres. If you asked users with the experience in both worlds, 
what do you think that the answer would be, which one is faster?


> There you go. PostgreSQL with its limited resources has to concentrate
> on the core system. It is not out to provide everything that Oracle
> does, but it is out to do the things it does better than everybody else.
> I believe that PostgreSQL is quite interested in what corporate users
> want. Think of in-core replication. But that does not imply that
> PostgreSQL can and wants to provide everything that corporate users ask
> for.

The case with hints shows a fatal flaw in the mentality of Postgres 
community. Quite frankly, if Postgres didn't make it to the prime time in 
16 years, I don't think it ever will. There are other options, from NoSQL 
databases like MongoDB to Firebird. I am waiting for Firebird 3.0 to 
assess how usable it will be. It's a major effort and from what I am 
reading about it, it definitely looks interesting.


> I am sure that you'll find the community helpful if you have questions.
> However, it will be pretty difficult to get that into core. You'd have
> to be patient at least.

Initially, I've asked for hints to only be put on the TODO list. At that 
time, that would have satisfied me. I was steamrolled. The "community" 
doesn't even want to talk about hints. It's a carnival of vanities, 
they've drawn the line in the sand and now there is no more discussion 
about that. Fine with me. 


> But as long as you can scratch your itch, that's a minor inconvenience.

I can, but not with Postgres. The absence of hints disqualifies Postgres 
for any further projects. And I wished for Postgres to succeed. I was 
hosting meetings and I was very enthusiastic about it. Until I touched 
the topic of hints. 


> As I mentioned, there *are* other things you can do. - You can examine
> the output of EXPLAIN ANALYZE yourself and figure out
>   where and why things go wrong.
> - You can influence statistics levels on a per-column basis. - You can
> influence query planning with a number of parameters.

Laurenz, you skipped all my explanations that humans are smarter than 
computer programs and that there must be the way for human to decide and 
override the decisions made by the optimizer. For a database to be 
viable, it must offer a tuning methodology, a finite set of steps that 
user can make to improve performance. I find posting to the mailing an 
unacceptable methodology. This was supposed to be a discussion on the 
technical merits of hints, yet there is no answer to the main points I've 
made. Let me repeat them once again:
- Humans are smarter than the optimizer and have better knowledge of
  the data. Therefore, they should have the possibility to override the 
  decisions made by the optimizer.
- Hints are needed, as the last resort, but they are. When there is no 
  other way of getting the desired performance, hints must be available
  as the method of the last resort.
- Database must offer a tuning methodology, things that user can do to 
  improve performance. You start with a SQL, you run explain analyze and 
  then there must be a series of steps to take. In PostgreSQL, there 
  isn't such a methodology.
- Such refusal to provide a method to override the optimizer decisions 
  reflects deep mistrust of the database developers toward their users.

The only reason for so stubbornly refusing hints is what I call the 
carnival of vanities, the line drawn in the sand by a group of 
developers, who have never been DBA's, and who will rather see their 
product to sink into obscurity than to give in to reasonable demands. At 
this point, I would like to hear one reason against hints. All I've heard 
so far is to be patient and that it's "deep in core", without any 
reasoning against having the option. For patience to make sense, someone 
must acknowledge that hints are necessary and put them on the TODO list. 
With pearls of wisdom like the one previously mentioned, patience makes 
no sense. A change in attitude is needed and I find that unlikely to 
happen.

> 
> That's ok, share your experiences.
> But if you want your opinion to be heard, refrain from insult and
> sweeping generalizations with no more proof than your authority.

I don't think that I am getting my message across here. I don't care 
about being heard by the Postgres community, that's way behind me. I've 
said what I've had, those who wanted to listen did so. Now, my goal is to 
try warning the Oracle users of the pitfalls and let them know that 
Postgres is essentially inadequate for the task. In other words, my 
message is not intended for the Postgres community at all, it is intended 
for the Oracle users. I've abandoned my hopes for Postgres. With Josh 
Berkus, Kevin Grittner,Joshua Drake, Tom Lane and Bruce Momjian, there 
isn't much hope for the project. Sure, it will live on, maybe for even 
another decade, but it will eventually sink into obscurity. Firebird 
looks promising, NoSQL databases seem like a good solution for the DW 
type use, and there are some other ideas, too. 

-- 
http://mgogala.byethost5.com

Back to comp.databases.postgresql | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

migrating oracle to postgres Username <user@domain.tld> - 2011-06-20 19:28 +0200
  Re: migrating oracle to postgres Mladen Gogala <no@email.here.invalid> - 2011-06-20 20:08 +0000
    Re: migrating oracle to postgres Username <user@domain.tld> - 2011-06-20 23:45 +0200
      Re: migrating oracle to postgres Mladen Gogala <no@email.here.invalid> - 2011-06-20 22:37 +0000
      Re: migrating oracle to postgres Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-21 02:01 +0000
  Re: migrating oracle to postgres "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-21 09:18 +0200
    Re: migrating oracle to postgres Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-21 13:38 +0000
      Is PostgreSQL good? (was: migrating oracle to postgres) "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-22 12:30 +0200
        Re: Is PostgreSQL good? Harry Tuttle <OTPXDAJCSJVU@spammotel.com> - 2011-06-22 13:15 +0200
        Re: Is PostgreSQL good? (was: migrating oracle to postgres) Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-22 12:02 +0000
          Re: Is PostgreSQL good? "M. Strobel" <sorry_no_mail_here@nowhere.dee> - 2011-06-22 21:52 +0200
            Re: Is PostgreSQL good? Mladen Gogala <no@email.here.invalid> - 2011-06-22 22:43 +0000
          Re: Is PostgreSQL good? (was: migrating oracle to postgres) "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-24 12:29 +0200
            Re: Is PostgreSQL good? Hans Castorp <REWYRLXHEGHO@spammotel.com> - 2011-06-24 12:57 +0200
              Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-24 14:53 +0200
              Re: Is PostgreSQL good? Mladen Gogala <no@email.here.invalid> - 2011-06-24 16:32 +0000
            Re: Is PostgreSQL good? (was: migrating oracle to postgres) Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-24 13:53 +0000
              Re: Is PostgreSQL good? (was: migrating oracle to postgres) "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-27 11:17 +0200
                Re: Is PostgreSQL good? (was: migrating oracle to postgres) Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-28 06:11 +0000
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-28 10:34 +0200
                Re: Is PostgreSQL good? Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-28 13:30 +0000
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-29 09:50 +0200
                Re: Is PostgreSQL good? Mladen Gogala <no@email.here.invalid> - 2011-06-29 21:38 +0000
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-06-30 09:14 +0200
                Re: Is PostgreSQL good? Mladen Gogala <mgogala@no.address.invalid> - 2011-06-30 13:22 +0000
            Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-06-30 07:05 -0700
              Re: Is PostgreSQL good? Mladen Gogala <no@email.here.invalid> - 2011-06-30 15:41 +0000
                Re: Is PostgreSQL good? Richard Kettlewell <rjk@greenend.org.uk> - 2011-06-30 16:49 +0100
                Re: Is PostgreSQL good? Mladen Gogala <no@email.here.invalid> - 2011-06-30 19:24 +0000
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-06-30 13:21 -0700
                Re: Is PostgreSQL good? Mladen Gogala <no@email.here.invalid> - 2011-06-30 22:42 +0000
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-07-01 11:57 -0700
                Re: Is PostgreSQL good? Mladen Gogala <gogala.mladen@gmail.com> - 2011-07-01 22:58 +0000
                Re: Is PostgreSQL good? Richard Kettlewell <rjk@greenend.org.uk> - 2011-07-01 14:07 +0100
                Re: Is PostgreSQL good? Mladen Gogala <gogala.mladen@gmail.com> - 2011-07-01 13:29 +0000
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-06-30 13:17 -0700
                Re: Is PostgreSQL good? Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-30 23:53 +0000
              Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-07-01 10:13 +0200
                Re: Is PostgreSQL good? Mladen Gogala <gogala.mladen@gmail.com> - 2011-07-01 13:38 +0000
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-07-01 16:58 +0200
                Re: Is PostgreSQL good? Mladen Gogala <no@email.here.invalid> - 2011-07-01 15:33 +0000
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-07-01 11:39 -0700
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-07-04 16:39 +0200
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-07-04 12:11 -0700
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-07-05 08:58 +0200
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-07-05 11:29 -0700
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-07-06 09:00 +0200
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-07-06 01:00 -0700
                Re: Is PostgreSQL good? "Laurenz Albe" <invite@spam.to.invalid> - 2011-07-07 09:05 +0200
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-07-07 00:34 -0700
                Re: Is PostgreSQL good? Matthew Woodcraft <mattheww@chiark.greenend.org.uk> - 2011-07-04 20:12 +0100
                Re: Is PostgreSQL good? Don Y <nowhere@here.com> - 2011-07-04 12:24 -0700
      Re: migrating oracle to postgres Robert Klemme <shortcutter@googlemail.com> - 2011-06-23 17:48 +0200
        Re: migrating oracle to postgres Mladen Gogala <no@email.here.invalid> - 2011-06-23 16:42 +0000
          Re: migrating oracle to postgres Mladen Gogala <no@email.here.invalid> - 2011-06-23 18:57 +0000
          Re: migrating oracle to postgres Robert Klemme <shortcutter@googlemail.com> - 2011-06-26 17:35 +0200
            Re: migrating oracle to postgres Mladen Gogala <gogala.mladen@gmail.com> - 2011-06-26 19:08 +0000
          Re: migrating oracle to postgres Robert Klemme <shortcutter@googlemail.com> - 2011-06-27 00:32 -0700
            Re: migrating oracle to postgres Mladen Gogala <no@email.here.invalid> - 2011-06-28 16:00 +0000
              Re: migrating oracle to postgres Robert Klemme <shortcutter@googlemail.com> - 2011-06-28 22:28 +0200

csiph-web