Обсуждение: I am being interviewed by OReilly
Hey, Oreilly and Assoc is interviewing me and they asked me two questions I don't have the answers to: When is 7.3 set to land? When is 8.0 set to land? I said, when there done, but they want a little more ;) Joshua Drake Co-Author Practical PostgreSQL Command Prompt, Inc. -- Creators of Mammoth PostgreSQL
"Joshua D. Drake" <jd@commandprompt.com> writes:
> Oreilly and Assoc is interviewing me and they asked me two questions I
> don't have the answers to:
> When is 7.3 set to land?
> When is 8.0 set to land?
7.3 will go beta at the end of August, barring major disasters.
As for final release, it's done when it's done --- the optimistic
schedule would be end of September, but we do not release by the
calendar. We release when we think the code is ready.
There is no plan anywhere that involves an 8.0; if anyone thinks
they know how many 7.* releases there will be, when 8.0 will be
out, or what will be in it, they are just blowing smoke. We have
a hard enough time seeing ahead to the next release...
regards, tom lane
On Wed, Jul 03, 2002 at 01:46:46PM -0700, Joshua D. Drake wrote:
>
> When is 7.3 set to land?
>
> When is 8.0 set to land?
As a matter of curiosity, what would constitute "8.0" as opposed to,
say, 7.4? (I know that 7.0 happened partly because a great whack of
new features went in, but I haven't found anything in the -hackers
archives to explain why the number change. Maybe it's just a phase
of the moon thing, or something.)
A
--
----
Andrew Sullivan 87 Mowat Avenue
Liberty RMS Toronto, Ontario Canada
<andrew@libertyrms.info> M6K 3E3
+1 416 646 3304 x110
> When is 8.0 set to land?
You might point out that every release is an 8.0 by the pathetic
standards now used by many or most products for labeling releases.
We take a perverse pride in versioning The Old Fashioned Way, perhaps to
an extreme ;)
- Thomas
Andrew Sullivan wrote: > On Wed, Jul 03, 2002 at 01:46:46PM -0700, Joshua D. Drake wrote: > > > > When is 7.3 set to land? > > > > When is 8.0 set to land? > > As a matter of curiosity, what would constitute "8.0" as opposed to, > say, 7.4? (I know that 7.0 happened partly because a great whack of > new features went in, but I haven't found anything in the -hackers > archives to explain why the number change. Maybe it's just a phase > of the moon thing, or something.) Actually, it was a wack of new features in 6.5 when we realized we had to up the version on the next release. I think multi-master replication would be an 8.0 item, and point-in-time recovery. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Andrew Sullivan <andrew@libertyrms.info> writes:
> As a matter of curiosity, what would constitute "8.0" as opposed to,
> say, 7.4? (I know that 7.0 happened partly because a great whack of
> new features went in, but I haven't found anything in the -hackers
> archives to explain why the number change. Maybe it's just a phase
> of the moon thing, or something.)
I remember quite a deal of argument about whether to call it 7.0 or 6.6;
we had started that cycle with the assumption that it would be called
6.6, and changed our minds near the end. Personally I'd have preferred
to stick the 7.* label on starting with the next release (actually
called 7.1) which had WAL and TOAST in it. That was really a
significant set of changes, both on the inside and outside.
You could make a fair argument that the upcoming 7.3 ought to be
called 8.0, because the addition of schema support will break an
awful lot of client-side code ;-). But I doubt we will do that.
regards, tom lane
Tom Lane wrote: > Andrew Sullivan <andrew@libertyrms.info> writes: > > As a matter of curiosity, what would constitute "8.0" as opposed to, > > say, 7.4? (I know that 7.0 happened partly because a great whack of > > new features went in, but I haven't found anything in the -hackers > > archives to explain why the number change. Maybe it's just a phase > > of the moon thing, or something.) > > I remember quite a deal of argument about whether to call it 7.0 or 6.6; > we had started that cycle with the assumption that it would be called > 6.6, and changed our minds near the end. Personally I'd have preferred > to stick the 7.* label on starting with the next release (actually > called 7.1) which had WAL and TOAST in it. That was really a > significant set of changes, both on the inside and outside. > > You could make a fair argument that the upcoming 7.3 ought to be > called 8.0, because the addition of schema support will break an > awful lot of client-side code ;-). But I doubt we will do that. Yes, the problem with incrementing on major features is that we would start to look like Emacs numbering fairly quickly. At some point, we may have to modify our name and start at 1.0 again. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Bruce Momjian wrote: > <snip> > > At some point, we may have to modify our name and start at 1.0 again. Heh Heh Heh Let's do the M$ trick and pick a name that everyone will confuse and assume it's us: "Standard SQL 1.0". So when people use the popularity question for deciding their database "what database does everyone else use? I just want the standard one..." We win. :) + Justin > > -- > Bruce Momjian | http://candle.pha.pa.us > pgman@candle.pha.pa.us | (610) 853-3000 > + If your life is a hard drive, | 830 Blythe Avenue > + Christ can be your backup. | Drexel Hill, Pennsylvania 19026 > > ---------------------------(end of broadcast)--------------------------- > TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) -- "My grandfather once told me that there are two kinds of people: those who work and those who take the credit. He told me to try to be in the first group; there was less competition there." - Indira Gandhi
On Thu, 4 Jul 2002, Bruce Momjian wrote: > Tom Lane wrote: > > Andrew Sullivan <andrew@libertyrms.info> writes: > > > As a matter of curiosity, what would constitute "8.0" as opposed to, > > > say, 7.4? (I know that 7.0 happened partly because a great whack of > > > new features went in, but I haven't found anything in the -hackers > > > archives to explain why the number change. Maybe it's just a phase > > > of the moon thing, or something.) > > > > I remember quite a deal of argument about whether to call it 7.0 or 6.6; > > we had started that cycle with the assumption that it would be called > > 6.6, and changed our minds near the end. Personally I'd have preferred > > to stick the 7.* label on starting with the next release (actually > > called 7.1) which had WAL and TOAST in it. That was really a > > significant set of changes, both on the inside and outside. > > > > You could make a fair argument that the upcoming 7.3 ought to be > > called 8.0, because the addition of schema support will break an > > awful lot of client-side code ;-). But I doubt we will do that. > > Yes, the problem with incrementing on major features is that we would > start to look like Emacs numbering fairly quickly. At 2.5years in v7.x, I think its going to be a long while before we start getting into the 20's :) > At some point, we may have to modify our name and start at 1.0 again. Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL' known, so let's change the name *just* so that we can start at 1.0 and face a new challenge of getting ppl to recognize the name?
"Marc G. Fournier" <scrappy@hub.org> writes:
>> At some point, we may have to modify our name and start at 1.0 again.
> Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL'
> known, so let's change the name *just* so that we can start at 1.0 and
> face a new challenge of getting ppl to recognize the name?
I've heard a number of people opine that we should go back to just plain
'Postgres', which is pronounceable by the uninitiate, and besides which
that's what we use informally most of the time. 'PostgreSQL' is about
as marketing-unfriendly a name as you could easily find...
I'd not be in favor of picking something new out of the blue, but I'd
pick 'Postgres' over 'PostgreSQL' if it were up to me.
regards, tom lane
Tom Lane wrote: > "Marc G. Fournier" <scrappy@hub.org> writes: > >> At some point, we may have to modify our name and start at 1.0 again. > > > Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL' > > known, so let's change the name *just* so that we can start at 1.0 and > > face a new challenge of getting ppl to recognize the name? > > I've heard a number of people opine that we should go back to just plain > 'Postgres', which is pronounceable by the uninitiate, and besides which > that's what we use informally most of the time. 'PostgreSQL' is about > as marketing-unfriendly a name as you could easily find... I personally agree. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Thu, 4 Jul 2002, Tom Lane wrote: > I'd not be in favor of picking something new out of the blue, but I'd > pick 'Postgres' over 'PostgreSQL' if it were up to me. As I recall the only real reason for the change was to emphasize that the query language had changed to SQL. Back in my young and naive days (probably early '95) I remember picking up Postgres, realizing it didn't use SQL as the query language, thinking, "How terrible!" and immediately dropping it for MySQL. (I'm older and wiser now, but it's too late--all the systems that let you use something less crappy than SQL are now gone. *Sigh*.) Anyway, I expect that others had the same experience, and thus something like that was required to get people who had previously dropped it to go back to it again. Now that QUEL or PostQUEL or whatever it was is long gone and fogotten (except maybe in certain CA-Unicenter shops), I see no reason we couldn't go back to "Postgres" now. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
Tom Lane wrote: > You could make a fair argument that the upcoming 7.3 ought to be > called 8.0, because the addition of schema support Star-schema support? will break an > awful lot of client-side code ;-).
On Thu, 4 Jul 2002, Tom Lane wrote: > "Marc G. Fournier" <scrappy@hub.org> writes: > >> At some point, we may have to modify our name and start at 1.0 again. > > > Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL' > > known, so let's change the name *just* so that we can start at 1.0 and > > face a new challenge of getting ppl to recognize the name? > > I've heard a number of people opine that we should go back to just plain > 'Postgres', which is pronounceable by the uninitiate, and besides which I can never figure this out ... what is so difficult about 'Postgres-Q-L'?
Can't figure it out cause you're a techie most of us that subscribe to these lists are... When I was doing tech support i had people asking for help with [these aren't typo's] Prostgre sequel Postresquirrel pgsql psql Progress etc... and alot of the folks i talked to were not newbies, they had been using it for a while. > -----Original Message----- > From: pgsql-general-owner@postgresql.org > [mailto:pgsql-general-owner@postgresql.org]On Behalf Of Marc G. Fournier > Sent: Friday, July 05, 2002 10:05 AM > To: Tom Lane > Cc: Bruce Momjian; Andrew Sullivan; pgsql-general@postgresql.org > Subject: Re: [GENERAL] I am being interviewed by OReilly > > > On Thu, 4 Jul 2002, Tom Lane wrote: > > > "Marc G. Fournier" <scrappy@hub.org> writes: > > >> At some point, we may have to modify our name and start at 1.0 again. > > > > > Ya, that's it ... we've only spent, what, 8 years now making > 'PostgreSQL' > > > known, so let's change the name *just* so that we can start at 1.0 and > > > face a new challenge of getting ppl to recognize the name? > > > > I've heard a number of people opine that we should go back to just plain > > 'Postgres', which is pronounceable by the uninitiate, and besides which > > I can never figure this out ... what is so difficult about 'Postgres-Q-L'? > > > > > > ---------------------------(end of broadcast)--------------------------- > TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) > > >
On Fri, 5 Jul 2002, Marc G. Fournier wrote: > On Thu, 4 Jul 2002, Tom Lane wrote: > > > "Marc G. Fournier" <scrappy@hub.org> writes: > > >> At some point, we may have to modify our name and start at 1.0 again. > > > > > Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL' > > > known, so let's change the name *just* so that we can start at 1.0 and > > > face a new challenge of getting ppl to recognize the name? > > > > I've heard a number of people opine that we should go back to just plain > > 'Postgres', which is pronounceable by the uninitiate, and besides which > > I can never figure this out ... what is so difficult about 'Postgres-Q-L'? Of course I got it completely round the wrong and thought Postgres was the latter of the two names. I've got to go change my documents now. -- Nigel J. Andrews Director --- Logictree Systems Limited Computer Consultants
Marc G. Fournier wrote: > On Thu, 4 Jul 2002, Tom Lane wrote: > > > "Marc G. Fournier" <scrappy@hub.org> writes: > > >> At some point, we may have to modify our name and start at 1.0 again. > > > > > Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL' > > > known, so let's change the name *just* so that we can start at 1.0 and > > > face a new challenge of getting ppl to recognize the name? > > > > I've heard a number of people opine that we should go back to just plain > > 'Postgres', which is pronounceable by the uninitiate, and besides which > > I can never figure this out ... what is so difficult about 'Postgres-Q-L'? Beats me, but when the Addison-Wesley publisher called to talk to me about doing a book, he called is Postgre. I knew we were in trouble. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
At 7/5/02 3:33 AM, Tom Lane wrote: >I've heard a number of people opine that we should go back to just plain >'Postgres', which is pronounceable by the uninitiate, and besides which >that's what we use informally most of the time. 'PostgreSQL' is about >as marketing-unfriendly a name as you could easily find... What's even worse is that when you do a search for a particular piece of information about PostgreSQL, you have to do the search twice, because you don't know whether the Web page or message that you're looking for would have referred to it as "Postgres" or "PostgreSQL". Heck, I can never even remember if the username created by the install is "postgres" or "postgresql". The fact that so many people don't use the full name "PostgreSQL" seems to me to be an indication in and of itself that something is wrong with it. And the fact that they use "Postgres" instead is an indication of what it should be called (the full official name could be "Postgres SQL Server"). I see the point of having "SQL" in the name -- it theoretically lets people know what it does, like "MySQL" -- but in practice, the disadvantages of "PostgreSQL" seem to far outweigh that benefit. ------------------------------------ Robert L Mathews, Tiger Technologies
Well, for one: An awful lot of people think the name is: "Postgre-SQL", or "Postgre-SEQUEL" If marketing matters (as a lot of people have been suggesting in recent threads), then better to have a non-ambiguous name that is easy to pronounce correctly. --- "Marc G. Fournier" <scrappy@hub.org> wrote: > I can never figure this out ... what is so difficult > about 'Postgres-Q-L'? > __________________________________________________ Do You Yahoo!? Sign up for SBC Yahoo! Dial - First Month Free http://sbc.yahoo.com
It is not intuitive. It is like pronouncing MySQL "mice -Q -L"
>>> "Marc G. Fournier" <scrappy@hub.org> 07/05/02 06:05AM >>>
On Thu, 4 Jul 2002, Tom Lane wrote:
> "Marc G. Fournier" <scrappy@hub.org> writes:
> >> At some point, we may have to modify our name and start at 1.0 again.
>
> > Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL'
> > known, so let's change the name *just* so that we can start at 1.0 and
> > face a new challenge of getting ppl to recognize the name?
>
> I've heard a number of people opine that we should go back to just plain
> 'Postgres', which is pronounceable by the uninitiate, and besides which
I can never figure this out ... what is so difficult about 'Postgres-Q-L'?
---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo@postgresql.org)
On Fri, Jul 05, 2002 at 01:50:05PM -0400, Bruce Momjian wrote:
>
> Beats me, but when the Addison-Wesley publisher called to talk to me
> about doing a book, he called is Postgre. I knew we were in trouble.
For what it's worth, the calling cards they printed for me here
(which I have used exactly 0 times) when I arrived have "Postgre SQL
Administrator" on them. And I corrected the typo 5 times and sent it
back for proof each time. They just didn't believe it, I guess.
Anyway, I never thing about it. But I _do_ tend to say "Postgres".
Possibly for the same reason that I find "GNU/Linux" to be too much
trouble.
Andrew "not to mention Micrsoft SQL Server 20,642" Sullivan
--
----
Andrew Sullivan 87 Mowat Avenue
Liberty RMS Toronto, Ontario Canada
<andrew@libertyrms.info> M6K 3E3
+1 416 646 3304 x110
Bruce Momjian dijo: > Marc G. Fournier wrote: > > I can never figure this out ... what is so difficult about 'Postgres-Q-L'? > > Beats me, but when the Addison-Wesley publisher called to talk to me > about doing a book, he called is Postgre. I knew we were in trouble. Lots of people here calls it "Postgre", even something like "Postgree" (Postgri in spanish). Others say "Postgre-S-Q-L". But the vast majority uses plain "Postgres". I have yet to meet somebody who says "Postgres-Q-L". I think the pronunciation is really counter-intuitive. Not that it's difficult to say; I think it's difficult to read. -- Alvaro Herrera (<alvherre[a]atentus.com>) "God is real, unless declared as int"
Sander Steffann wrote: > > Beats me, but when the Addison-Wesley publisher called to talk to me > > about doing a book, he called is Postgre. I knew we were in trouble. > > I know someone who does that too (even after telling him it is NOT Postgre a > LOT of times)... I just don't know where people get that idea! I didn't write it correctly. He said "Postgray" I think or "Postgree". Both made me feel a little ill. The strange thing is that until PostgreSQL got popular, no one really said the word, we just wrote it, and my editor has a "PostgreSQL" string macro so I don't even type it anymore, but when you start to talk to people, it does become a problem. The idea of calling it "Postgres SQL Server" has merit because it is so close to what we already have, just an added 's' and a space. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Fri, 5 Jul 2002, Jeff Eckermann wrote: > Well, for one: > > An awful lot of people think the name is: > "Postgre-SQL", or > "Postgre-SEQUEL" > > If marketing matters (as a lot of people have been > suggesting in recent threads), then better to have a > non-ambiguous name that is easy to pronounce > correctly. Post-gres-Q-L ... again, that is difficult to pronounce in what way? Or is it just too much work to educate ppl? I'm sorry, but I have alot of ppl that ask me about 'Postgres', and the first thing I do is explain to them what *Postgres* was, and how to pronounce Post-gres-Q-L ... personally, I've been saying it for so long now that it just rolls off the tongue *shrug*
On Fri, 5 Jul 2002, Ian Harding wrote: > It is not intuitive. It is like pronouncing MySQL "mice -Q -L" > > >>> "Marc G. Fournier" <scrappy@hub.org> 07/05/02 06:05AM >>> > On Thu, 4 Jul 2002, Tom Lane wrote: > > > "Marc G. Fournier" <scrappy@hub.org> writes: > > >> At some point, we may have to modify our name and start at 1.0 again. > > > > > Ya, that's it ... we've only spent, what, 8 years now making 'PostgreSQL' > > > known, so let's change the name *just* so that we can start at 1.0 and > > > face a new challenge of getting ppl to recognize the name? > > > > I've heard a number of people opine that we should go back to just plain > > 'Postgres', which is pronounceable by the uninitiate, and besides which > > I can never figure this out ... what is so difficult about 'Postgres-Q-L'? From my point of view ... its not that it difficult to pronounce, its just too much work for ppl to bother correcting, so they propogate the incorrect name ... it wouldn't be a problem if ppl took the two minutes to teach ppl instead of just letting it slip ...
On Sat, 6 Jul 2002, Marc G. Fournier wrote: > Post-gres-Q-L ... again, that is difficult to pronounce in what way? With the accent on which sylable? Rich :-)
On Fri, 5 Jul 2002, Rich Shepard wrote:
> On Sat, 6 Jul 2002, Marc G. Fournier wrote:
>
> > Post-gres-Q-L ... again, that is difficult to pronounce in what way?
>
> With the accent on which sylable?
I never really thought about it ... but you tell me:
http://www2.ca.postgresql.org/postgresql.mp3
I don't really hear an accent on any of the syllables in there, but I've
been known to be tone deaf too :)
On Sat, 6 Jul 2002, Marc G. Fournier wrote: > On Fri, 5 Jul 2002, Rich Shepard wrote: > > > On Sat, 6 Jul 2002, Marc G. Fournier wrote: > > > > > Post-gres-Q-L ... again, that is difficult to pronounce in what way? > > > > With the accent on which sylable? > > I never really thought about it ... but you tell me: > > http://www2.ca.postgresql.org/postgresql.mp3 > > I don't really hear an accent on any of the syllables in there, but I've > been known to be tone deaf too :) I can't say I've had any trouble thinking how to say PostgreSQL but I don't think it rolls off the tongue as easily as SQL Server for instance. Although I think that name is a good name using Postgres SQL Server would enable people to shorten it to just the Postgres part while still clearly identifying what it is they are on about, I'm ignoring the issue of more experienced people remembering/knowing about just plain Postgres. Plus it has the added benefit of being similar to another product, i.e. Microsoft SQL Server, which people already say. Sometimes, okay so most times, this is shortened to SQL Server and we'd just be reversing this emphasis. Anyway, I do quite like the PostgreSQL, even if I've had to go back and capitalise that S twice now, I just thought I'd point out the blindingly obvious. -- Nigel J. Andrews Director --- Logictree Systems Limited Computer Consultants
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> The idea of calling it "Postgres SQL Server" has merit because it is so
> close to what we already have, just an added 's' and a space.
... and a M$ trademark violation suit, just waiting to happen whenever
M$ decides we are big enough to be a threat.
Stay far far away from any name including "SQL Server".
(I still like plain "Postgres" though.)
regards, tom lane
On Sat, 6 Jul 2002, Marc G. Fournier wrote: > I never really thought about it ... but you tell me: > > http://www2.ca.postgresql.org/postgresql.mp3 > > I don't really hear an accent on any of the syllables in there, but I've > been known to be tone deaf too :) No sound card or speakers here, only the built-in squeeker that comes with all units. In fact, I was just teasing, anyway, because the thread on the name and its pronounciation had taken on such a long and serious life. From a linguistic perspective, however, I tend to pronounce it with the accent on the second syllable. Ciao, Rich Dr. Richard B. Shepard, President Applied Ecosystem Services, Inc. (TM) 2404 SW 22nd Street | Troutdale, OR 97060-1247 | U.S.A. + 1 503-667-4517 (voice) | + 1 503-667-8863 (fax) | rshepard@appl-ecosys.com http://www.appl-ecosys.com
Tom Lane wrote: > Bruce Momjian <pgman@candle.pha.pa.us> writes: > >>The idea of calling it "Postgres SQL Server" has merit because it is so >>close to what we already have, just an added 's' and a space. > > > ... and a M$ trademark violation suit, just waiting to happen whenever > M$ decides we are big enough to be a threat. > > Stay far far away from any name including "SQL Server". IIRC SQL Server (or "SQL-server" to be exact) is just a definition from the SQL standard. I doubt whether Microsoft can monopolize something like that (the SQL standard was there first right?). Of course I am making a pretty big assumption about the sanity of the actual law :) Whether it is really worthwhile to invest resources into that is an entirely different matter. Personally, I am happy with PostgreSQL. Jochem
Jochem van Dieten <jochemd@oli.tudelft.nl> writes:
>> Stay far far away from any name including "SQL Server".
> IIRC SQL Server (or "SQL-server" to be exact) is just a definition from
> the SQL standard. I doubt whether Microsoft can monopolize something
> like that (the SQL standard was there first right?).
Microsoft has managed to make "Windows" into a trademark, even though
it's by rights a generic term. Yes, I know the name of their product
is really "Microsoft Windows", but just try calling something "Windows"
and see what happens ...
Even if we avoid any trademark problems, "Postgres SQL Server" just
seems way too much like a me-too name. It *will* cause confusion
with Microsoft's product.
regards, tom lane
Marc G. Fournier wrote: > On Fri, 5 Jul 2002, Jeff Eckermann wrote: > > > Well, for one: > > > > An awful lot of people think the name is: > > "Postgre-SQL", or > > "Postgre-SEQUEL" > > > > If marketing matters (as a lot of people have been > > suggesting in recent threads), then better to have a > > non-ambiguous name that is easy to pronounce > > correctly. > > Post-gres-Q-L ... again, that is difficult to pronounce in what way? Or I think the problem is that PostgreSQL is both a name and an acronym, mixed into single word. Postgres is a name, SQL is an acronym, PostgreSQL is both. That is where people get confused. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Bruce Momjian <pgman@candle.pha.pa.us> writes:
>> Post-gres-Q-L ... again, that is difficult to pronounce in what way? Or
> I think the problem is that PostgreSQL is both a name and an acronym,
> mixed into single word. Postgres is a name, SQL is an acronym,
> PostgreSQL is both. That is where people get confused.
The problem is that the typography makes it look like the split should
be "Postgre / SQL". I can see exactly why people think the name portion
is "Postgre" --- it's not at all apparent that the "S" is part of both
parts of the word, until you've been told.
Had we capitalized the name like "PostgresQL" maybe the correct
pronunciation would be more obvious.
regards, tom lane
> I think the problem is that PostgreSQL is both a name and an acronym,
> mixed into single word. Postgres is a name, SQL is an acronym,
> PostgreSQL is both. That is where people get confused.
thats the point..
i think "postgres" sounds better :)
infact i have posted , a couple of times, to pgsql-general@postgres.org :p
^_^
--
Varun
------
@n=(544290696690,305106661574,116357),$b=16,@c=' .JPacehklnorstu'=~
/./g;for$n(@n){map{$h=int$n/$b**$_;$n-=$b**$_*$h;$c[@c]=$h}c(0..9);
push@p,map{$c[$_]}@c[c($b..$#c)];$#c=$b-1}print@p;sub'c{reverse @_}
Tom Lane wrote: > Bruce Momjian <pgman@candle.pha.pa.us> writes: > >> Post-gres-Q-L ... again, that is difficult to pronounce in what way? Or > > > I think the problem is that PostgreSQL is both a name and an acronym, > > mixed into single word. Postgres is a name, SQL is an acronym, > > PostgreSQL is both. That is where people get confused. > > The problem is that the typography makes it look like the split should > be "Postgre / SQL". I can see exactly why people think the name portion > is "Postgre" --- it's not at all apparent that the "S" is part of both > parts of the word, until you've been told. > > Had we capitalized the name like "PostgresQL" maybe the correct > pronunciation would be more obvious. Yep, there is no other word that has an acronym part and a word part, and where capitalization not match in the two parts. It is almost a recipe for confusion. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Tom Lane wrote: > Bruce Momjian <pgman@candle.pha.pa.us> writes: > >> Post-gres-Q-L ... again, that is difficult to pronounce in what way? Or > > > I think the problem is that PostgreSQL is both a name and an acronym, > > mixed into single word. Postgres is a name, SQL is an acronym, > > PostgreSQL is both. That is where people get confused. > > The problem is that the typography makes it look like the split should > be "Postgre / SQL". I can see exactly why people think the name portion > is "Postgre" --- it's not at all apparent that the "S" is part of both > parts of the word, until you've been told. > > Had we capitalized the name like "PostgresQL" maybe the correct > pronunciation would be more obvious. No one pronounces MySQL as Mice-QL. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Sat, 6 Jul 2002, Tom Lane wrote: > Bruce Momjian <pgman@candle.pha.pa.us> writes: > > The idea of calling it "Postgres SQL Server" has merit because it is so > > close to what we already have, just an added 's' and a space. > > ... and a M$ trademark violation suit, just waiting to happen whenever > M$ decides we are big enough to be a threat. > > Stay far far away from any name including "SQL Server". > > (I still like plain "Postgres" though.) Nobody to date, I don't believe, has jumped down ppls throat for informally calling it Postres .. the "formal" name is PostgreSQL .. nothing stop's ppl from using Postgres in 'conversation' though ... I personaly use PgSQL more often then not, since I find ppl seem to be able to spell it, while alot of ppl that I've dealt with have a problem with spelling the Postgres part of PostgreSQL for some reason ...
Marc G. Fournier wrote: > On Sat, 6 Jul 2002, Tom Lane wrote: > > > Bruce Momjian <pgman@candle.pha.pa.us> writes: > > > The idea of calling it "Postgres SQL Server" has merit because it is so > > > close to what we already have, just an added 's' and a space. > > > > ... and a M$ trademark violation suit, just waiting to happen whenever > > M$ decides we are big enough to be a threat. > > > > Stay far far away from any name including "SQL Server". > > > > (I still like plain "Postgres" though.) > > Nobody to date, I don't believe, has jumped down ppls throat for > informally calling it Postres .. the "formal" name is PostgreSQL .. > nothing stop's ppl from using Postgres in 'conversation' though ... I > personaly use PgSQL more often then not, since I find ppl seem to be able I think that is because PgSQL is a full acronym, rather than a mixed word/acronym combination, though actually if Pg were a word, they would say pu-gu-SQL. The 'my' in MySQL is a word so they say it that way. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Sat, 6 Jul 2002, Tom Lane wrote:
> Had we capitalized the name like "PostgresQL" maybe the correct
> pronunciation would be more obvious.
^^^^
Remove this word and you'd be dead on. Perhaps the simplest resolution is
to do just this.
Rich
Rich Shepard wrote: > On Sat, 6 Jul 2002, Tom Lane wrote: > > > Had we capitalized the name like "PostgresQL" maybe the correct > > pronunciation would be more obvious. > ^^^^ > > Remove this word and you'd be dead on. Perhaps the simplest resolution is > to do just this. What we really need then is Postgres-Q-L. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Sat, 6 Jul 2002, Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > The idea of calling it "Postgres SQL Server" has merit because it is so
> > close to what we already have, just an added 's' and a space.
>
> ... and a M$ trademark violation suit, just waiting to happen whenever
> M$ decides we are big enough to be a threat.
>
> Stay far far away from any name including "SQL Server".
>
> (I still like plain "Postgres" though.)
Here in Russia, most people use "Postgres".
>
> regards, tom lane
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster
>
>
Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83
On Sat, Jul 06, 2002 at 02:30:17PM -0400, Bruce Momjian wrote: > Marc G. Fournier wrote: > > On Sat, 6 Jul 2002, Tom Lane wrote: > > > > > Bruce Momjian <pgman@candle.pha.pa.us> writes: > > > > The idea of calling it "Postgres SQL Server" has merit because it is so > > > > close to what we already have, just an added 's' and a space. > > > > > > ... and a M$ trademark violation suit, just waiting to happen whenever > > > M$ decides we are big enough to be a threat. > > > > > > Stay far far away from any name including "SQL Server". > > > > > > (I still like plain "Postgres" though.) > > > > Nobody to date, I don't believe, has jumped down ppls throat for > > informally calling it Postres .. the "formal" name is PostgreSQL .. > > nothing stop's ppl from using Postgres in 'conversation' though ... I > > personaly use PgSQL more often then not, since I find ppl seem to be able > > I think that is because PgSQL is a full acronym, rather than a mixed > word/acronym combination, though actually if Pg were a word, they would > say pu-gu-SQL. The 'my' in MySQL is a word so they say it that way. I often hear it pronounced Pig-Squeal, which is memorable, but may not give the right impression. Cheers, Steve
On Sat, 6 Jul 2002, Bruce Momjian wrote:
> What we really need then is Postgres-Q-L.
That, or PostgreS-Q-L would both make clear the pronounciation and retain
the differentiation from the former Postgres. Businesses change names and
logos quite successfully. No reason not to make such a change if it resolves
everyone's concerns.
Rich
Dr. Richard B. Shepard, President
Applied Ecosystem Services, Inc. (TM)
2404 SW 22nd Street | Troutdale, OR 97060-1247 | U.S.A.
+ 1 503-667-4517 (voice) | + 1 503-667-8863 (fax) | rshepard@appl-ecosys.com
http://www.appl-ecosys.com
Hi,
> I didn't write it correctly. He said "Postgray" I think or "Postgree".
> Both made me feel a little ill. The strange thing is that until
> PostgreSQL got popular, no one really said the word, we just wrote it,
> and my editor has a "PostgreSQL" string macro so I don't even type it
> anymore, but when you start to talk to people, it does become a problem.
>
> The idea of calling it "Postgres SQL Server" has merit because it is so
> close to what we already have, just an added 's' and a space.
Is this under discussion? I do think that "PostgreSQL" is a good
name but does cause a lot of confusion. I usually call it just "Postgres".
I don't see the necessity of having "SQL" in it's name, although "Postgres
SQL" is also interesting.
[]'s
Ricardo.
Maybe not so obvious when people start asking what this "QL"
means. ;-)
[]'s
Ricardo.
On Sat, 6 Jul 2002, Tom Lane wrote:
> > I think the problem is that PostgreSQL is both a name and an acronym,
> > mixed into single word. Postgres is a name, SQL is an acronym,
> > PostgreSQL is both. That is where people get confused.
>
> The problem is that the typography makes it look like the split should
> be "Postgre / SQL". I can see exactly why people think the name portion
> is "Postgre" --- it's not at all apparent that the "S" is part of both
> parts of the word, until you've been told.
>
> Had we capitalized the name like "PostgresQL" maybe the correct
> pronunciation would be more obvious.
>
> regards, tom lane
G'Day > Bruce Momjian <pgman@candle.pha.pa.us> writes: > No one pronounces MySQL as Mice-QL. Wanna bet? I have heard some atrocious and perverted attempts at both MySQL (micesql, my sequel, etc) and PostgreSQL, but then again I am an Aussie and live is Orstralia (.AU) where we speak Strine (Australian). The people who have the most problems are non-technical. Our previous Business Development Manager almost choked on PostgreSQL when he first came across it. He was reading for a presentation and did not ask first. So when it came to say the word it was 'Postgrrrr....mumble'. Luckily it was not in front of a customer :-) Regards, Kym Farnik (mailto:kym@recalldesign.com) -- Recall Design http://www.recalldesign.com 53 Gilbert Street, Adelaide, South Australia 5000 Direct: (61-8) 8217 0556 Fax: (61-8) 8217 0555 Mobile: 0438 014 007
G'Day :-) > Bruce Momjian <pgman@candle.pha.pa.us> writes: > No one pronounces MySQL as Mice-QL. Wanna bet? I have heard some atrocious and perverted attempts at both MySQL (micesql, my sequel, etc) and PostgreSQL, but then again I am an Aussie and live is Orstralia (.AU) where we speak Strine (Australian). Which brings up the old Sequel vs S Q L. (Dont go here) The people who have the most problems are non-technical. Our previous Business Development Manager almost choked on PostgreSQL when he first came across it. He was reading for a presentation and did not ask first. So when it came to say the word it was 'Postgrrrr....mumble'. Luckily it was not in front of a customer :-) Anyway... Love it or hate it: Postgres is what we use internally and Postgres Q L in front of customers. Regards, Kym Farnik (mailto:kym@recalldesign.com) -- Recall Design http://www.recalldesign.com 53 Gilbert Street, Adelaide, South Australia 5000 Direct: (61-8) 8217 0556 Fax: (61-8) 8217 0555 Mobile: 0438 014 007
On Sat, 6 Jul 2002, Ricardo Junior wrote: > On Sat, 6 Jul 2002, Tom Lane wrote: > > > Had we capitalized the name like "PostgresQL" maybe the correct > > pronunciation would be more obvious. > > Maybe not so obvious when people start asking what this "QL" > means. ;-) That's obvious. Since QL means "query language," "Postgres QL" would refer to the old, QUEL-derived query language that Postgres used before it was ripped out and replaced with SQL, right? "Postgres" is simple, people use it anyway, and everybody now knows that Postgres uses SQL instead of its own query language now, so I think it would be a very good to just switch back to to that. With the demise of Great Bridge, we even have the postgres.org domain name free for this now. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
Hi Tom: > You could make a fair argument that the upcoming 7.3 ought to be > called 8.0, because the addition of schema support will break an > awful lot of client-side code ;-). But I doubt we will do that. Hmm. As it happens, I've written an awful lot of client-side code. Can you elaborate on what will break, or point me to a resource that lays it out? -- sgl ======================================================= Steve Lane Vice President Chris Moyer Consulting, Inc. 833 West Chicago Ave Suite 203 Voice: (312) 433-2421 Email: slane@fmpro.com Fax: (312) 850-3930 Web: http://www.fmpro.com =======================================================
> That's obvious. Since QL means "query language," "Postgres QL" would
> refer to the old, QUEL-derived query language that Postgres used before
> it was ripped out and replaced with SQL, right?
>
> "Postgres" is simple, people use it anyway, and everybody now knows that
> Postgres uses SQL instead of its own query language now, so I think it
> would be a very good to just switch back to to that. With the demise of
> Great Bridge, we even have the postgres.org domain name free for this
> now.
So obvious for us, so enigmatic for non-technical people. Also,
"PostgresQL" could very well be understood as "another kind of SQL", as
people could read "Postgres Query Language". I agree that "Postgres" is a
better approach for what is desired. I still would like to know if this is
really under discussion, I mean, if such change would be possible someday.
[]'s
Ricardo.
> -----Original Message----- > From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > Sent: Saturday, July 06, 2002 8:20 AM > To: Bruce Momjian > Cc: Sander Steffann; Marc G. Fournier; Andrew Sullivan; > pgsql-general@postgresql.org > Subject: Re: [GENERAL] I am being interviewed by OReilly > > > Bruce Momjian <pgman@candle.pha.pa.us> writes: > > The idea of calling it "Postgres SQL Server" has merit > because it is so > > close to what we already have, just an added 's' and a space. > > ... and a M$ trademark violation suit, just waiting to happen whenever > M$ decides we are big enough to be a threat. > > Stay far far away from any name including "SQL Server". > > (I still like plain "Postgres" though.) http://zdnet.com.com/2100-1104-862516.html (march 2002 article) 'However, in a preliminary ruling issued late Friday, U.S. District Judge John Coughenour said the suit raised "serious questions" about whether the word "Windows" is entitled to trademark protection.' http://zdnet.com.com/2100-1107-893895.html (april 2002) 'Here's why: An elementary principle of trademark law says that generic terms cannot be protected. No company could obtain trademark rights to the word "computer" to describe what we all know as a computer, for example.' My Google searching didn't turn up anything suggesting that this case had ever been resolved-- does anybody know differently? -ron > > regards, tom lane > > > > ---------------------------(end of > broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster > >
Hi Ricardo, suga@netbsd.com.br wrote: > <snip> > I agree that "Postgres" is a > better approach for what is desired. I still would like to know if this is > really under discussion, I mean, if such change would be possible someday. It's interesting to know what people's opinions are, and I don't feel the desire for a simpler name to be wrong. However, we all have taken a *lot* of time and effort to get the name "PostgreSQL" recognised, and we should continue on doing this. Nowdays I'm finding it very unusual to see new articles and publications going online and getting it wrong, meaning that although there are legacy documents out there refering to "Postgres", most of the new stuff online is calling it the proper "PostgreSQL". Lets stick with "PostgreSQL", simply because we are a solid project with a good foundation, and we don't need to introduce uncertainties in people's minds by changing the "Official" name of our project every few years. "PostgreSQL" works, people recognise it, and its our brand alone. :) Regards and best wishes, Justin Clift > []'s > Ricardo. > > ---------------------------(end of broadcast)--------------------------- > TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) -- "My grandfather once told me that there are two kinds of people: those who work and those who take the credit. He told me to try to be in the first group; there was less competition there." - Indira Gandhi
At 7/7/02 10:08 PM, Ron Snyder wrote: >> Stay far far away from any name including "SQL Server". >> >> (I still like plain "Postgres" though.) > >http://zdnet.com.com/2100-1104-862516.html (march 2002 article) >'However, in a preliminary ruling issued late Friday, U.S. District Judge >John Coughenour said the suit raised "serious questions" about whether the >word "Windows" is entitled to trademark protection.' > >http://zdnet.com.com/2100-1107-893895.html (april 2002) >'Here's why: An elementary principle of trademark law says that generic >terms cannot be protected. No company could obtain trademark rights to the >word "computer" to describe what we all know as a computer, for example.' > >My Google searching didn't turn up anything suggesting that this case had >ever been resolved-- does anybody know differently? Don't know about that, but for what it's worth, a TESS trademark search indicates that the phrase "SQL Server" is a registered trademark of... Sybase(!). I'm surprised it was allowed to be registered, actually; it seems merely descriptive. One could certainly make an argument that since Microsoft also uses the phrase, it's generic, but making such an argument in court would probably take more money than all of us have put together :-) ------------------------------------ Robert L Mathews, Tiger Technologies
On Sun, 7 Jul 2002, Robert L Mathews wrote: > Don't know about that, but for what it's worth, a TESS trademark search > indicates that the phrase "SQL Server" is a registered trademark of... > Sybase(!). > > I'm surprised it was allowed to be registered, actually; it seems merely > descriptive. One could certainly make an argument that since Microsoft > also uses the phrase, it's generic.... They may have the right to use it; Microsoft SQL Server is based on Sybase code. You didn't think they'd come up with a product that good all on their own, did you? :-) cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
Hello all, I'm still trying to track down my very odd periodic pauses/hangs in PostgreSQL 7.2.1. I've localized it to what seems to be the "recycled transaction log file" lines in the log file. Whenever this happens, a whole bunch of queries which were "on hold" (just sitting there, as can be seen in pg_stat_activity, when they usually execute in fractions of a second) come back to life and finish very quickly. Unfortunately, PostgreSQL doesn't seem to log when it starts doing this recycling, only when it's done. However, it seems to be taking about 1.5 minutes (yes, around 90 seconds) to do this recycling on about sixteen of these WAL files at a time. (Deduction from the logs from the application that uses the database.) I currently have about 102 of these WAL files (I don't mind; I have 50 gigs set aside for pg_xlog). My postgresql.conf settings are: WAL_FILES = 48 WAL_BUFFERS = 16 CHECKPOINT_SEGMENTS = 30 With this, during my heavy load period, I get those 16 WAL recycling messages every 6.5 minutes. During heavy vacuuming, the recycling happens every 3 minutes, and that was my goal (no more than every three minutes, per Bruce Momjian's PDF on tuning). My server specs: Dual P4 Xeon 2.4 8gb RAM RAID-1 drive for pg_xlog - running ext3 RAID-5 drive dedicated to PostgreSQL for everything else - running ext3 Debian 3.0 (woody) kernel 2.4.18 Some questions: 1) Is there any known bad interactions with ext3fs and PostgreSQL? My hardware vendor (Pogo Linux, recommended) seemed to suggest that ext3fs has problems in multi-threading. 2) Any ideas on how to get it to log more info on WAL usage? 3) Which process in PostgreSQL should I attach to using gdb to check out this WAL stuff? Putting my application on hold for 1.5 minutes out of every 6.5 is of course very bad... I'm stumped. Any ideas are welcome; I am willing to provide any additional information and run any other tests. Thanks, Doug
Hi there,
> It's interesting to know what people's opinions are, and I don't feel
> the desire for a simpler name to be wrong.
> However, we all have taken a *lot* of time and effort to get the name
> "PostgreSQL" recognised, and we should continue on doing this. Nowdays
> I'm finding it very unusual to see new articles and publications going
> online and getting it wrong, meaning that although there are legacy
> documents out there refering to "Postgres", most of the new stuff online
> is calling it the proper "PostgreSQL".
I can't agree more with you. I do think that some people being
confused about saying "PostgreSQL" is NOT enough reason to change a name
that, as you said, was recognised by means of hard work. That's also why I
asked if this was really under discussion. The "desired", to be more
clear, referred to a name that would solve this "issue", and not
substitute the present name. Anyway, the problem saying PostgreSQL does
exist. Someone even mentioned "PostgresQL", a name that isn't really to my
taste. :-)
> Lets stick with "PostgreSQL", simply because we are a solid project with
> a good foundation, and we don't need to introduce uncertainties in
> people's minds by changing the "Official" name of our project every few
> years. "PostgreSQL" works, people recognise it, and its our brand
> alone. :)
Well said, and I'm glad to see that even the developers gave their
opinions on the subject. This kind of talk is usually restricted to people
without envolvement with the related project.
[]'s
Ricardo.
On Mon, 8 Jul 2002 suga@netbsd.com.br wrote: > > > That's obvious. Since QL means "query language," "Postgres QL" would > > refer to the old, QUEL-derived query language that Postgres used before > > it was ripped out and replaced with SQL, right? > > > > "Postgres" is simple, people use it anyway, and everybody now knows that > > Postgres uses SQL instead of its own query language now, so I think it > > would be a very good to just switch back to to that. With the demise of > > Great Bridge, we even have the postgres.org domain name free for this > > now. > > > So obvious for us, so enigmatic for non-technical people. Also, > "PostgresQL" could very well be understood as "another kind of SQL", as > people could read "Postgres Query Language". I agree that "Postgres" is a > better approach for what is desired. I still would like to know if this is > really under discussion, I mean, if such change would be possible someday. No, no more then switching to GPL is an option ...
On Fri, 5 Jul 2002, Marc G. Fournier wrote: > > I've heard a number of people opine that we should go back to just plain > > 'Postgres', which is pronounceable by the uninitiate, and besides which > > I can never figure this out ... what is so difficult about 'Postgres-Q-L'? Nothing. We even have audio files on the website so there is no question on how to pronounce it. Some folks just aren't happy unless they can change what doesn't need changing is all. I guess it's their way of contributing. Vince. -- ========================================================================== Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net 56K Nationwide Dialup from $16.00/mo at Pop4 Networking Online Campground Directory http://www.camping-usa.com Online Giftshop Superstore http://www.cloudninegifts.com ==========================================================================
> Beats me, but when the Addison-Wesley publisher called to talk to me > about doing a book, he called is Postgre. I knew we were in trouble. I know someone who does that too (even after telling him it is NOT Postgre a LOT of times)... I just don't know where people get that idea! Sander.
> Bruce Momjian <pgman@candle.pha.pa.us> writes: > > The idea of calling it "Postgres SQL Server" has merit because it is so > > close to what we already have, just an added 's' and a space. > > ... and a M$ trademark violation suit, just waiting to happen whenever > M$ decides we are big enough to be a threat. > > Stay far far away from any name including "SQL Server". Then we just leave out the 'Server' :-) > (I still like plain "Postgres" though.) If you make it 'Postgres SQL' you have it all. Sure, everybody will call it 'Postgres' (like a lot of them do now). But you still have the 'PostgreSQL' feeling... Just my feeling though... Sander
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > Nothing. We even have audio files on the website so there is > no question on how to pronounce it. Some folks just aren't > happy unless they can change what doesn't need changing is > all. I guess it's their way of contributing. I don't understand some of the animosity I have seen towards what I consider a positive change. It obviously does need changing, otherwise so many people would not be discussing it. An audio file on the website is not going to help people who are reading the word inside text somewhere. The very fact that such an audio file even needs to exist should be telling us something. It was created because people were having a hard time pronouncing it. Saying that people can pronounce it properly because there is an audio file smacks of circular logic. > However, we all have taken a *lot* of time and effort to get > the name "PostgreSQL" recognised, and we should continue on > doing this. Nowdays I'm finding it very unusual to see new > articles and publications going online and getting it wrong, > meaning that although there are legacy documents out there > refering to "Postgres", most of the new stuff online is > calling it the proper "PostgreSQL". Sorry, but the shortcut "Postgres" is alive and kicking. Look anywhere, even in the mailing lists. The default user is "postgres" not "postgresql". Another point is that postgres fits in the 8.3 naming schema, which is not really used anymore, but does serve as a fairly good rule of thumb. A product name should be 8 letters or less (Linux, Apache, Windows, Sybase, Oracle, Ingres, Sybase, MySQL, Apache, tinydns, sendmail, qmail, iptables, etc.) (Microsoft products are an exception of course, but they have the marketing power to name something "throatwobblermangrove" and still have the masses purchase it :) I might agree somewhat with the "don't break tradition" argument if there had been a concerted effort from the start to *dissuade* people from using the word "postgres", but as far as I can tell, most people involved with Postgre[sS][QL] don't really care which one is being used, and all other things being equal, tend to simplify it to the shorter form, especially when talking out loud. I've given presentations on Postgres to technical and non-technical people, and always have to throw in a section at the start about the name - how to pronounce the "long form", how it came about, and to not worry about which term I use within the presentation, since they are synonymous. I'd rather spend that time extolling some of the better virtues of the product, rather than trying to explain why we smushed a word and an acronym together into something awkward to pronounce. Perhaps that's the best way to put it: the word is not difficult to pronounce, but it is awkward. Greg Sabino Mullane greg@turnstep.com PGP Key: 0x14964AC8 200207081020 -----BEGIN PGP SIGNATURE----- Comment: http://www.turnstep.com/pgp.html iD8DBQE9KaCJvJuQZxSWSsgRAloRAKD4tbIititzKXI08kEpAFSkeT/YrgCfcyDa 0UOzfOknsIEi7B1kfdTFFGU= =ADDF -----END PGP SIGNATURE-----
Doug Fields <dfields-pg-general@pexicom.com> writes:
> I'm still trying to track down my very odd periodic pauses/hangs in
> PostgreSQL 7.2.1.
> I've localized it to what seems to be the "recycled transaction log file"
> lines in the log file. Whenever this happens, a whole bunch of queries
> which were "on hold" (just sitting there, as can be seen in
> pg_stat_activity, when they usually execute in fractions of a second) come
> back to life and finish very quickly.
Hm. That would be from the tail end of a checkpoint operation. If you
do a CHECKPOINT command manually, do you get a similar hangup?
What are the stuck queries doing exactly? Can you attach to a few of the
stuck backends with gdb and get stack traces?
> 1) Is there any known bad interactions with ext3fs and PostgreSQL?
Dunno. The CHECKPOINT would probably create a significant number of
disk write requests, followed by a sync() request. If that could
monopolize an ext3 filesystem for a long time, perhaps that would
explain your problem. But I haven't heard similar complaints before.
What do you have shared_buffers set to?
regards, tom lane
Steve Lane <slane@fmpro.com> writes:
>> You could make a fair argument that the upcoming 7.3 ought to be
>> called 8.0, because the addition of schema support will break an
>> awful lot of client-side code ;-). But I doubt we will do that.
> Hmm. As it happens, I've written an awful lot of client-side code. Can you
> elaborate on what will break, or point me to a resource that lays it out?
Anything that looks at the system catalogs is likely to have some
trouble; the notion that there is at most one pg_class row named 'foo',
for example, will fall down.
regards, tom lane
At 11:57 AM 7/8/2002, Tom Lane wrote:
>Doug Fields <dfields-pg-general@pexicom.com> writes:
> > I'm still trying to track down my very odd periodic pauses/hangs in
> > PostgreSQL 7.2.1.
> > I've localized it to what seems to be the "recycled transaction log file"
> > lines in the log file. Whenever this happens, a whole bunch of queries
> > which were "on hold" (just sitting there, as can be seen in
> > pg_stat_activity, when they usually execute in fractions of a second) come
> > back to life and finish very quickly.
>
>Hm. That would be from the tail end of a checkpoint operation. If you
>do a CHECKPOINT command manually, do you get a similar hangup?
In fact, yes. (I never did a CHECKPOINT ever before now.) It took less time
(probably because I did it around half way through the 6.5-minute WAL
recycling period), but had almost identical effects. pg_stat_activity
showed a whole bunch of queries stacked up, and vmstat exhibited the same
pathology: the other postgres processes blocked during the checkpoint (it's
pretty plain in the vmstat output below).
procs memory swap io system
cpu
r b w swpd free buff cache si so bi bo in cs us
sy id
0 0 0 0 3851828 121732
3913552 0 0 2 79 56 62 49 2 49
1 0 0 0 3850724 121732
3913576 0 0 0 1171 214 1023 4 0 95
1 0 0 0 3851544 121732
3913576 0 0 0 266 179 148 1 0 99
0 0 0 0 3851780 121732
3913600 0 0 0 1407 266 364 4 1 95
0 2 1 0 3843020 121732
3913616 0 0 0 4182 195 186 2 4 94
0 2 3 0 3841608 121732
3913616 0 0 0 2722 197 191 0 1 99
0 2 3 0 3840984 121732
3913616 0 0 0 2774 232 298 1 1 99
0 2 3 0 3840056 121732
3913616 0 0 0 2739 200 204 1 1 99
0 2 3 0 3839280 121732
3913616 0 0 0 1634 183 201 1 0 99
0 2 4 0 3839424 121732
3913616 0 0 0 4098 202 143 0 1 99
0 2 4 0 3837672 121732
3913616 0 0 0 2710 194 170 1 1 99
0 2 3 0 3836992 121732
3913616 0 0 0 2139 156 82 0 0 99
0 2 4 0 3827844 121732
3913632 0 0 0 5362 170 3260 3 1 95
0 2 4 0 3828592 121732
3913632 0 0 0 2134 143 53 0 1 99
1 1 0 0 3830892 121744
3913668 0 0 0 4926 152 6087 8 3 89
0 0 0 0 3850360 121744 3913772 0 0 0 7851 610
16106 61 5 34
0 0 0 0 3850360 121744
3913772 0 0 0 20 189 284 1 0 99
0 0 0 0 3850360 121744
3913772 0 0 0 14 147 157 0 0 100
>What are the stuck queries doing exactly? Can you attach to a few of the
>stuck backends with gdb and get stack traces?
The stuck queries are of the following nature, typically either:
1) UPDATE x SET somedate=now() WHERE primarykey IN (1,2,3,4...300 to 450 IDs)
2) DELETE FROM x WHERE key1 = 1 AND key2 IN (1,2,3,4...300 to 450 IDs)
In my pg_stats display below you can see some of them.
Here is a stack trace. I did "where" about every second during the "pause"
and received the same stack trace. This is on PID 3456 per the
pg_stat_activity listing below. After things clear up, I also did a stack
trace; it's blocked on recv, presumably waiting for more commands to come
down the socket. (I tried a few other PIDs with similar stack traces, all
stuck on the semop call.)
(gdb) c
Continuing.
Program received signal SIGINT, Interrupt.
0x4028913f in semop () from /lib/libc.so.6
(gdb) where
#0 0x4028913f in semop () from /lib/libc.so.6
#1 0x08106dec in IpcSemaphoreLock ()
#2 0x0810bce1 in LWLockAcquire ()
#3 0x0808433b in XLogInsert ()
#4 0x08079033 in _bt_doinsert ()
#5 0x08078ea9 in _bt_doinsert ()
#6 0x080786ed in _bt_doinsert ()
#7 0x0807be9d in btinsert ()
#8 0x08152822 in OidFunctionCall5 ()
#9 0x08077cf8 in index_insert ()
#10 0x080c674a in ExecInsertIndexTuples ()
#11 0x080c373e in ExecutorEnd ()
#12 0x080c3374 in ExecutorEnd ()
#13 0x080c2797 in ExecutorRun ()
#14 0x081104de in ProcessQuery ()
#15 0x0810ed70 in pg_exec_query_string ()
#16 0x0810fd5e in PostgresMain ()
#17 0x080f6d4e in ClosePostmasterPorts ()
#18 0x080f669f in ClosePostmasterPorts ()
#19 0x080f5882 in PostmasterMain ()
#20 0x080f5391 in PostmasterMain ()
#21 0x080d4e18 in main ()
#22 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) c
Continuing.
And then later...
(gdb) c
Continuing.
Program received signal SIGINT, Interrupt.
0x402887f2 in recv () from /lib/libc.so.6
(gdb) where
#0 0x402887f2 in recv () from /lib/libc.so.6
#1 0x080d42bc in StreamClose ()
#2 0x080d430d in pq_getbyte ()
#3 0x0810e7c8 in HandleFunctionRequest ()
#4 0x0810e837 in HandleFunctionRequest ()
#5 0x0810fc5e in PostgresMain ()
#6 0x080f6d4e in ClosePostmasterPorts ()
#7 0x080f669f in ClosePostmasterPorts ()
#8 0x080f5882 in PostmasterMain ()
#9 0x080f5391 in PostmasterMain ()
#10 0x080d4e18 in main ()
#11 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) c
Continuing.
> > 1) Is there any known bad interactions with ext3fs and PostgreSQL?
>
>Dunno. The CHECKPOINT would probably create a significant number of
>disk write requests, followed by a sync() request. If that could
>monopolize an ext3 filesystem for a long time, perhaps that would
>explain your problem. But I haven't heard similar complaints before.
>
>What do you have shared_buffers set to?
pexicast_lg=# show shared_buffers;
NOTICE: shared_buffers is 65536
SHOW VARIABLE
I believe this is about 512 megs. I have 8 gigs RAM on this server and
tried 65536 and before it 32768. Either one seems to work fine - I didn't
notice any significant performance difference yet between them.
See below for pg_stat_activity...
I appreciate the help!
Cheers,
Doug
pexicast_lg=# select * from pg_stat_activity where datname='pexicast_lg';
select now();
datid | datname | procpid | usesysid | usename
|
current_query
---------+-------------+---------+----------+---------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
3926366 | pexicast_lg | 3393 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(463196,463216,463236,463256,463276,463296,461176,461196,461216,461236,461256,461276,461296,461316,461336,461356,461376,461396,461416,461436,461456,461476,461496,461516,461536,461556,461576,46
3926366 | pexicast_lg | 3412 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(425066,425086,425106,425126,425146,425166,425186,425206,425226,425246,425266,425286,425306,425326,425346,425366,425386,425406,425426,425446,425466,425486,425506,425526,425546,425566,425586,42
3926366 | pexicast_lg | 3443 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(442093,442113,442133,442153,442173,442193,442213,442233,442253,442273,442293,442313,442333,442353,442373,442393,442413,442433,442453,442473,442493,442513,442533,442553,442573,442593,442613,44
3926366 | pexicast_lg | 3444 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(438107,438127,438147,438167,438187,438207,438227,438247,436147,436167,436187,436207,436227,436247,436267,436287,436307,436327,436347,436367,436387,436407,436427,436447,436467,436487,436507,43
3926366 | pexicast_lg | 3446 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(429470,429490,429510,429530,429550,429570,429590,429610,429630,429650,429670,429690,429710,429730,429750,429770,429790,429810,429830,429850,429870,429890,429910,429930,427830,427850,427870,42
3926366 | pexicast_lg | 3447 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(463658,463678,463698,463718,463738,463758,463778,463798,463818,463838,463858,463878,463898,463918,460378,461838,461858,461878,461898,461918,461938,461958,461978,461998,462018,462038,462058,46
3926366 | pexicast_lg | 3448 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(407502,407522,407542,407562,407582,407602,407622,407642,407662,407682,407702,407722,407742,407762,407782,407802,407822,407842,407862,407882,405802,405822,405842,405862,405882,405902,405922,40
3926366 | pexicast_lg | 3449 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(475717,475737,475757,475777,475797,475817,475837,475857,475877,475897,475917,475937,475957,475977,475997,476017,476037,476057,476077,476097,476117,476137,476157,474077,474097,474117,474137,47
3926366 | pexicast_lg | 3450 | 100 | tomcat | <IDLE>
3926366 | pexicast_lg | 3451 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(438472,438492,438512,438532,438552,438572,438592,438612,438632,438652,438672,436572,436592,436612,436632,436652,436672,436692,436712,436732,436752,436772,436792,436812,436832,436852,436872,43
3926366 | pexicast_lg | 3452 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(415401,415421,415441,415461,415481,415501,415521,415541,415561,415581,415601,415621,415641,415661,415681,413581,413601,413621,413641,413661,413681,413701,413721,413741,413761,413781,413801,41
3926366 | pexicast_lg | 3453 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(442908,442928,442948,442968,440868,440888,440908,440928,440948,440968,440988,441008,441028,441048,441068,441088,441108,441128,441148,441168,441188,441208,441228,441248,441268,441288,441308,44
3926366 | pexicast_lg | 3454 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(410524,410544,410564,410584,410604,410624,410644,410664,410684,410704,410724,410744,408644,408664,408684,408704,408724,408744,408764,408784,408804,408824,408844,408864,408884,408904,408924,40
3926366 | pexicast_lg | 3455 | 100 | tomcat | <IDLE>
3926366 | pexicast_lg | 3456 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(451115,451135,451155,451175,451195,451215,451235,451255,451275,451295,451315,451335,451355,451375,451395,451415,451435,451455,451475,451495,451515,451535,451555,449475,449495,449515,449535,44
3926366 | pexicast_lg | 3457 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(408325,408345,408365,408385,408405,408425,408445,408465,408485,408505,408525,408545,406465,406485,406505,406525,406545,406565,406585,406605,406625,406645,406665,406685,406705,406725,406745,40
3926366 | pexicast_lg | 3458 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(397080,397100,397120,397140,397160,397180,397200,397220,397240,397260,397280,397300,397320,397340,397360,397380,397400,397420,397440,397460,397480,397500,397520,397540,397560,397580,397600,39
3926366 | pexicast_lg | 3459 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(476459,476479,476499,476519,476539,476559,476579,476599,476619,476639,476659,476679,476699,476719,476739,476759,474659,474679,474699,474719,474739,474759,474779,474799,474819,474839,474859,47
3926366 | pexicast_lg | 3460 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(411003,411023,411043,411063,411083,411103,411123,411143,411163,411183,411203,411223,411243,411263,411283,411303,411323,411343,411363,411383,411403,411423,411443,411463,411483,411503,411523,41
3926366 | pexicast_lg | 3461 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(432589,432609,432629,432649,432669,432689,432709,432729,432749,432769,432789,432809,432829,432849,432869,432889,432909,430809,430829,430849,430869,430889,430909,430929,430949,430969,430989,43
3926366 | pexicast_lg | 3462 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(470654,470674,470694,470714,470734,470754,470774,470794,470814,470834,470854,470874,470894,470914,470934,470954,470974,470994,471014,471034,471054,471074,471094,471114,471134,471154,471174,47
3926366 | pexicast_lg | 3497 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(440551,440571,440591,440611,440631,440651,440671,440691,440711,440731,440751,440771,440791,440811,440831,440851,440871,440891,440911,440931,440951,440971,440991,441011,441031,441051,441071,44
3926366 | pexicast_lg | 4235 | 101 | dfields | <IDLE>
(23 rows)
Hi Tom, all,
>Also, could you do the checkpoint manually and get a stack trace from
>that backend while others are hung up?
Yes, see below.
>directly try to acquire ControlFileLock. In any case it's hard to
>credit that the recycling process could take 90 seconds to rename a
>dozen or so files. If you have a gdb attached to a process doing a
>manual checkpoint, it would be fairly easy to see how long
>MoveOfflineLogs() runs. (Set a breakpoint at its start, when control
>reaches the breakpoint issue "fin" and see how long it takes to come
>back.)
Two things:
1) Remounted my ext3 filesystems as ext2 to rule out an ext3fs related
problem; the problems persist, so it's probably not an ext3 thing
2) Doing as you suggest, attaching to a /usr/lib/postgresql/bin/postgres
process from which I run a manual checkpoint results in the following behavior:
I was able to set a breakpoint at CreateCheckPoint - gdb never found a
MoveOfflineLogs for me to set a breakpoint.
Doing a single step from the breakpoint after the CreateCheckPoint takes
quite a few moments (perhaps 10-30 seconds, but I didn't use a stopwatch).
During this time control-C has no effect, and when it does take effect, it
leaves me at the displayed location below.
Single stepping from mdsync() to smgrsync() also seems to take a few
seconds (perhaps 3).
My GDB output is below. The checkpoint seems to take 42 seconds to complete:
pexicast_lg=# select now(); checkpoint; select now();
now
-------------------------------
2002-07-08 13:57:57.918766-04
(1 row)
CHECKPOINT
now
-------------------------------
2002-07-08 13:58:39.790787-04
(1 row)
I was wondering if it could be the sync function call, but I sit there on
the server and type sync until I'm blue in the face and they seem to run so
fast I don't even notice any delay.
Also, I can't seem to recompile PostgreSQL because Debian can't find a
tclConfig.sh.
Thanks,
Doug
(gdb) c
Continuing.
Breakpoint 1, 0x08087ae5 in CreateCheckPoint ()
(gdb) where
#0 0x08087ae5 in CreateCheckPoint ()
#1 0x08111066 in ProcessUtility ()
#2 0x0810ecc5 in pg_exec_query_string ()
#3 0x0810fd5e in PostgresMain ()
#4 0x080f6d4e in ClosePostmasterPorts ()
#5 0x080f669f in ClosePostmasterPorts ()
#6 0x080f5882 in PostmasterMain ()
#7 0x080f5391 in PostmasterMain ()
#8 0x080d4e18 in main ()
#9 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) n
Single stepping until exit from function CreateCheckPoint,
which has no line number information.
Program received signal SIGINT, Interrupt.
0x40282967 in sync () from /lib/libc.so.6
(gdb) where
#0 0x40282967 in sync () from /lib/libc.so.6
#1 0x0810d167 in mdsync ()
#2 0x0810de5f in smgrsync ()
#3 0x081036d8 in FlushBufferPool ()
#4 0x08087d13 in CreateCheckPoint ()
#5 0x08111066 in ProcessUtility ()
#6 0x0810ecc5 in pg_exec_query_string ()
#7 0x0810fd5e in PostgresMain ()
#8 0x080f6d4e in ClosePostmasterPorts ()
#9 0x080f669f in ClosePostmasterPorts ()
#10 0x080f5882 in PostmasterMain ()
#11 0x080f5391 in PostmasterMain ()
#12 0x080d4e18 in main ()
#13 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) break MoveOfflineLogs
Function "MoveOfflineLogs" not defined.
(gdb) break MoveOfflineLog
Function "MoveOfflineLog" not defined.
(gdb) break MoveOfflineLogs()
Function "MoveOfflineLogs()" not defined.
(gdb) s
Single stepping until exit from function sync,
which has no line number information.
0x0810d167 in mdsync ()
(gdb) s
Single stepping until exit from function mdsync,
which has no line number information.
0x0810de5f in smgrsync ()
(gdb) s
Single stepping until exit from function smgrsync,
which has no line number information.
0x081036d8 in FlushBufferPool ()
(gdb) where
#0 0x081036d8 in FlushBufferPool ()
#1 0x08087d13 in CreateCheckPoint ()
#2 0x08111066 in ProcessUtility ()
#3 0x0810ecc5 in pg_exec_query_string ()
#4 0x0810fd5e in PostgresMain ()
#5 0x080f6d4e in ClosePostmasterPorts ()
#6 0x080f669f in ClosePostmasterPorts ()
#7 0x080f5882 in PostmasterMain ()
#8 0x080f5391 in PostmasterMain ()
#9 0x080d4e18 in main ()
#10 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) s
Single stepping until exit from function FlushBufferPool,
which has no line number information.
0x08087d13 in CreateCheckPoint ()
(gdb) where
#0 0x08087d13 in CreateCheckPoint ()
#1 0x08111066 in ProcessUtility ()
#2 0x0810ecc5 in pg_exec_query_string ()
#3 0x0810fd5e in PostgresMain ()
#4 0x080f6d4e in ClosePostmasterPorts ()
#5 0x080f669f in ClosePostmasterPorts ()
#6 0x080f5882 in PostmasterMain ()
#7 0x080f5391 in PostmasterMain ()
#8 0x080d4e18 in main ()
#9 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) s
Single stepping until exit from function CreateCheckPoint,
which has no line number information.
0x08085ac4 in XLogFlush ()
(gdb) where
#0 0x08085ac4 in XLogFlush ()
#1 0x08111066 in ProcessUtility ()
#2 0x0810ecc5 in pg_exec_query_string ()
#3 0x0810fd5e in PostgresMain ()
#4 0x080f6d4e in ClosePostmasterPorts ()
#5 0x080f669f in ClosePostmasterPorts ()
#6 0x080f5882 in PostmasterMain ()
#7 0x080f5391 in PostmasterMain ()
#8 0x080d4e18 in main ()
#9 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) s
Single stepping until exit from function XLogFlush,
which has no line number information.
Program received signal SIGINT, Interrupt.
0x402887f2 in recv () from /lib/libc.so.6
(gdb) where
#0 0x402887f2 in recv () from /lib/libc.so.6
#1 0x080d42bc in StreamClose ()
#2 0x080d430d in pq_getbyte ()
#3 0x0810e7c8 in HandleFunctionRequest ()
#4 0x0810e837 in HandleFunctionRequest ()
#5 0x0810fc5e in PostgresMain ()
#6 0x080f6d4e in ClosePostmasterPorts ()
#7 0x080f669f in ClosePostmasterPorts ()
#8 0x080f5882 in PostmasterMain ()
#9 0x080f5391 in PostmasterMain ()
#10 0x080d4e18 in main ()
#11 0x401d114f in __libc_start_main () from /lib/libc.so.6
(gdb) c
Continuing.
Hello all,
I've some very interesting data collected from vmstat on the above problem.
I've now recorded the "whole cycle" on a vmstat log, using "vmstat 5" on my
server. I will annotate it here: This shows activity, context switches, and
"running, blocked and waiting" processes every 5 seconds.
procs memory swap io system
cpu
r b w swpd free buff cache si so bi bo in cs us
sy id
0 0 0 0 3858124 121692
3909904 0 0 0 863 234 298 4 0 96
1 0 0 0 3858116 121692
3909912 0 0 0 508 185 270 3 0 97
1 0 0 0 3857744 121692
3909936 0 0 0 940 224 487 6 0 94
0 0 0 0 3858076 121692
3909952 0 0 0 619 228 232 3 0 97
0 0 0 0 3858056 121696
3909968 0 0 0 582 194 193 3 0 97
0 0 0 0 3858024 121696
3910000 0 0 0 1036 249 276 5 0 95
0 0 0 0 3858024 121696
3910000 0 0 0 26 172 248 0 0 100
0 0 0 0 3858000 121696
3910024 0 0 0 902 243 458 4 0 96
1 0 0 0 3857732 121696
3910032 0 0 0 293 159 131 1 0 99
0 0 0 0 3857968 121696
3910056 0 0 0 742 206 186 3 0 96
0 0 0 0 3857960 121696
3910064 0 0 0 334 165 143 1 0 99
0 0 0 0 3857952 121696
3910072 0 0 0 320 164 153 1 0 98
0 0 0 0 3857936 121696
3910088 0 0 0 514 215 262 3 0 97
Up to this point everything is cranking along quite nicely. With 8gb of
RAM, you'll notice that it's extraordinarily rare for me to ever get any
"blocks in" - everything is always cached by the kernel. On the other hand,
the "blocks out" is fairly consistently high, as I have about 240 rows or
so being updated per second (usually in batches of 300-450) - some being
deleted and others having a single column updated.
Now, in the next 5 second analysis, something happens. I believe the WAL
logger begins its "duty cycle" and starts recycling log files. Doing
"select * from pg_stat_activity" shows all my normally very fast queries
stuck not being processed. Indeed, the vmstat log below continues to tell
the story: blocks out increases greatly, lots of processes (postgresql
since that's the _only_ thing that runs on this server other than system
daemons like syslogd) become blocked or waiting, etc.
At the end, the coup de grace comes: a flurry of context switches and tons
of blocks out. This isn't actually accurate enough for me to say which of
these two is the cause, but it's probably some combination of both:
1) The WAL finishes with a huge amount of I/O all at once
2) All the waiting queries (mostly UPDATE and DELETE) all of a sudden run
simultaneously
But either way, at the end of it, the write amounts jump sky high and so do
the context switches... Note that at 5 seconds per entry, this means about
1:15 (one and a quarter minutes) of time is lost, and as I said previously,
this happens approximately every six and a half minutes.
procs memory swap io system
cpu
r b w swpd free buff cache si so bi bo in cs us
sy id
0 3 1 0 3852388 121696
3910088 0 0 0 4257 205 151 1 7 92
0 2 3 0 3850944 121696
3910088 0 0 0 3000 222 264 0 1 99
0 2 3 0 3849920 121696
3910088 0 0 0 2643 176 132 0 1 99
0 2 3 0 3843964 121696
3910096 0 0 0 4035 208 1480 2 1 97
0 2 4 0 3840304 121696
3910096 0 0 0 3030 197 165 1 1 98
0 2 4 0 3839696 121696
3910096 0 0 0 2780 179 124 0 1 99
0 2 4 0 3838168 121696
3910096 0 0 0 2914 169 82 1 1 98
0 2 4 0 3838840 121696
3910096 0 0 0 2847 163 55 0 1 99
0 2 4 0 3839484 121696
3910096 0 0 0 2631 154 46 0 1 99
0 2 3 0 3840208 121696
3910096 0 0 0 2864 163 63 0 1 99
0 2 4 0 3840896 121696
3910096 0 0 0 2730 157 60 0 1 99
0 2 4 0 3841448 121696
3910096 0 0 0 2792 155 70 0 1 99
0 2 4 0 3841756 121696
3910096 0 0 0 2824 167 88 0 1 99
0 2 2 0 3841624 121696
3910096 0 0 0 2863 167 89 0 1 99
0 2 2 0 3841656 121696
3910096 0 0 0 2840 149 59 0 1 99
0 2 2 0 3841636 121696
3910096 0 0 0 3076 164 82 0 1 99
0 1 0 0 3827484 121696
3910128 0 0 0 4582 144 5577 6 2 92
5 0 0 0 3847556 121696 3910252 0 0 0 9406 207
13897 38 6 56
Hereafter, things settle down to normal. Actually, they settle down to
super-normal for a few moments, as all the threads (there are usually about
20-40 or more) now have another group of 300-450 rows to process, and won't
hit the database for about a minute to 90 seconds.
procs memory swap io system
cpu
r b w swpd free buff cache si so bi bo in cs us
sy id
0 0 0 0 3857684 121696
3910252 0 0 0 242 600 6071 26 2 72
0 0 0 0 3857684 121696
3910252 0 0 0 14 147 152 0 0 100
0 0 0 0 3857684 121696
3910252 0 0 0 14 144 145 0 0 100
0 0 0 0 3857672 121696
3910256 0 0 0 43 106 21 0 0 100
I welcome additional thoughts. My "SHOW ALL" command output is below.
Cheers,
Doug
pexicast_lg=# show all;
NOTICE: enable_seqscan is on
NOTICE: enable_indexscan is on
NOTICE: enable_tidscan is on
NOTICE: enable_sort is on
NOTICE: enable_nestloop is on
NOTICE: enable_mergejoin is on
NOTICE: enable_hashjoin is on
NOTICE: ksqo is off
NOTICE: geqo is on
NOTICE: tcpip_socket is on
NOTICE: ssl is off
NOTICE: fsync is on
NOTICE: silent_mode is off
NOTICE: log_connections is on
NOTICE: log_timestamp is on
NOTICE: log_pid is on
NOTICE: debug_print_query is off
NOTICE: debug_print_parse is off
NOTICE: debug_print_rewritten is off
NOTICE: debug_print_plan is off
NOTICE: debug_pretty_print is off
NOTICE: show_parser_stats is off
NOTICE: show_planner_stats is off
NOTICE: show_executor_stats is off
NOTICE: show_query_stats is off
NOTICE: stats_start_collector is on
NOTICE: stats_reset_on_server_start is on
NOTICE: stats_command_string is on
NOTICE: stats_row_level is off
NOTICE: stats_block_level is off
NOTICE: trace_notify is off
NOTICE: hostname_lookup is off
NOTICE: show_source_port is off
NOTICE: sql_inheritance is on
NOTICE: australian_timezones is off
NOTICE: fixbtree is on
NOTICE: password_encryption is off
NOTICE: transform_null_equals is off
NOTICE: geqo_threshold is 11
NOTICE: geqo_pool_size is 0
NOTICE: geqo_effort is 1
NOTICE: geqo_generations is 0
NOTICE: geqo_random_seed is -1
NOTICE: deadlock_timeout is 1000
NOTICE: syslog is 2
NOTICE: max_connections is 256
NOTICE: shared_buffers is 65536
NOTICE: port is 5432
NOTICE: unix_socket_permissions is 511
NOTICE: sort_mem is 65536
NOTICE: vacuum_mem is 65536
NOTICE: max_files_per_process is 1000
NOTICE: debug_level is 0
NOTICE: max_expr_depth is 10000
NOTICE: max_fsm_relations is 100
NOTICE: max_fsm_pages is 10000
NOTICE: max_locks_per_transaction is 64
NOTICE: authentication_timeout is 60
NOTICE: pre_auth_delay is 0
NOTICE: checkpoint_segments is 30
NOTICE: checkpoint_timeout is 300
NOTICE: wal_buffers is 16
NOTICE: wal_files is 48
NOTICE: wal_debug is 0
NOTICE: commit_delay is 0
NOTICE: commit_siblings is 5
NOTICE: effective_cache_size is 625000
NOTICE: random_page_cost is 4
NOTICE: cpu_tuple_cost is 0.01
NOTICE: cpu_index_tuple_cost is 0.001
NOTICE: cpu_operator_cost is 0.0025
NOTICE: geqo_selection_bias is 2
NOTICE: default_transaction_isolation is read committed
NOTICE: dynamic_library_path is $libdir
NOTICE: krb_server_keyfile is unset
NOTICE: syslog_facility is LOCAL0
NOTICE: syslog_ident is postgres
NOTICE: unix_socket_group is unset
NOTICE: unix_socket_directory is unset
NOTICE: virtual_host is unset
NOTICE: wal_sync_method is fdatasync
NOTICE: DateStyle is ISO with European conventions
NOTICE: Time zone is unset
NOTICE: TRANSACTION ISOLATION LEVEL is READ COMMITTED
NOTICE: Current client encoding is 'SQL_ASCII'
NOTICE: Current server encoding is 'SQL_ASCII'
NOTICE: Seed for random number generator is unavailable
SHOW VARIABLE
At 02:36 AM 7/8/2002, Doug Fields wrote:
>Hello all,
>
>I'm still trying to track down my very odd periodic pauses/hangs in
>PostgreSQL 7.2.1.
>
>I've localized it to what seems to be the "recycled transaction log file"
>lines in the log file. Whenever this happens, a whole bunch of queries
>which were "on hold" (just sitting there, as can be seen in
>pg_stat_activity, when they usually execute in fractions of a second) come
>back to life and finish very quickly.
>
>Unfortunately, PostgreSQL doesn't seem to log when it starts doing this
>recycling, only when it's done.
>
>However, it seems to be taking about 1.5 minutes (yes, around 90 seconds)
>to do this recycling on about sixteen of these WAL files at a time.
>(Deduction from the logs from the application that uses the database.) I
>currently have about 102 of these WAL files (I don't mind; I have 50 gigs
>set aside for pg_xlog). My postgresql.conf settings are:
>
>WAL_FILES = 48
>WAL_BUFFERS = 16
>CHECKPOINT_SEGMENTS = 30
>
>With this, during my heavy load period, I get those 16 WAL recycling
>messages every 6.5 minutes. During heavy vacuuming, the recycling happens
>every 3 minutes, and that was my goal (no more than every three minutes,
>per Bruce Momjian's PDF on tuning).
>
>My server specs:
>Dual P4 Xeon 2.4
>8gb RAM
>RAID-1 drive for pg_xlog - running ext3
>RAID-5 drive dedicated to PostgreSQL for everything else - running ext3
>Debian 3.0 (woody) kernel 2.4.18
>
>Some questions:
>
>1) Is there any known bad interactions with ext3fs and PostgreSQL? My
>hardware vendor (Pogo Linux, recommended) seemed to suggest that ext3fs
>has problems in multi-threading.
>2) Any ideas on how to get it to log more info on WAL usage?
>3) Which process in PostgreSQL should I attach to using gdb to check out
>this WAL stuff?
>
>Putting my application on hold for 1.5 minutes out of every 6.5 is of
>course very bad... I'm stumped. Any ideas are welcome; I am willing to
>provide any additional information and run any other tests.
>
>Thanks,
>
>Doug
Doug Fields <dfields-pg-general@pexicom.com> writes:
> I was able to set a breakpoint at CreateCheckPoint - gdb never found a
> MoveOfflineLogs for me to set a breakpoint.
Grumble. It's a static routine, and I imagine its name has been totally
stripped from the executable.
> Also, I can't seem to recompile PostgreSQL because Debian can't find a
> tclConfig.sh.
Oliver, what's an appropriate way to rebuild PG with debug symbols from
the Debian package?
regards, tom lane
Doug Fields <dfields-pg-general@pexicom.com> writes:
> Here is a stack trace. I did "where" about every second during the "pause"
> and received the same stack trace. This is on PID 3456 per the
> pg_stat_activity listing below. After things clear up, I also did a stack
> trace; it's blocked on recv, presumably waiting for more commands to come
> down the socket. (I tried a few other PIDs with similar stack traces, all
> stuck on the semop call.)
Hmm. I don't think I entirely believe that stack trace --- at least
some of the claimed call paths are impossible. Would it be too much
trouble to rebuild PG with --enable-debug and try again?
Also, could you do the checkpoint manually and get a stack trace from
that backend while others are hung up?
I am considering the possibility that the other backends are hung trying
to get ControlFileLock, which the checkpointer will acquire while
recycling xlog file segments --- but if your stack trace is accurate and
representative then that's not the problem because XLogInsert doesn't
directly try to acquire ControlFileLock. In any case it's hard to
credit that the recycling process could take 90 seconds to rename a
dozen or so files. If you have a gdb attached to a process doing a
manual checkpoint, it would be fairly easy to see how long
MoveOfflineLogs() runs. (Set a breakpoint at its start, when control
reaches the breakpoint issue "fin" and see how long it takes to come
back.)
regards, tom lane
From someone whose DSN used to look like: PostgreSQL I can tell you that P O S T G R E S Q L means *nothing* to people who are new to it. However, one can clearly see SQL at the end, and recognise that as a valid acronym, so assume that the prefix, and hence the product identifier, is POSTGRE That's where even educated people (perhaps even ESPECIALLY educated people) get the idea. Terry Fielder Network Engineer Great Gulf Homes / Ashton Woods Homes terry@greatgulfhomes.com > -----Original Message----- > From: pgsql-general-owner@postgresql.org > [mailto:pgsql-general-owner@postgresql.org]On Behalf Of > Sander Steffann > Sent: Friday, July 05, 2002 6:44 PM > To: Bruce Momjian; Marc G. Fournier > Cc: Tom Lane; Andrew Sullivan; pgsql-general@postgresql.org > Subject: Re: [GENERAL] I am being interviewed by OReilly > > > > Beats me, but when the Addison-Wesley publisher called to talk to me > > about doing a book, he called is Postgre. I knew we were > in trouble. > > I know someone who does that too (even after telling him it > is NOT Postgre a > LOT of times)... I just don't know where people get that idea! > > Sander. > > > > > > > ---------------------------(end of > broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org >
Bruce Momjian wrote: > > Marc G. Fournier wrote: > > On Sat, 6 Jul 2002, Tom Lane wrote: > > > > > Bruce Momjian <pgman@candle.pha.pa.us> writes: > > > > The idea of calling it "Postgres SQL Server" has merit because it is so > > > > close to what we already have, just an added 's' and a space. > > > > > > ... and a M$ trademark violation suit, just waiting to happen whenever > > > M$ decides we are big enough to be a threat. > > > > > > Stay far far away from any name including "SQL Server". > > > > > > (I still like plain "Postgres" though.) > > > > Nobody to date, I don't believe, has jumped down ppls throat for > > informally calling it Postres .. the "formal" name is PostgreSQL .. > > nothing stop's ppl from using Postgres in 'conversation' though ... I > > personaly use PgSQL more often then not, since I find ppl seem to be able > > I think that is because PgSQL is a full acronym, rather than a mixed > word/acronym combination, though actually if Pg were a word, they would > say pu-gu-SQL. The 'my' in MySQL is a word so they say it that way. Wouldn't that be "PiggySeeQuel" ? Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #================================================== JanWieck@Yahoo.com #
At 03:26 PM 7/8/2002, Tom Lane wrote: >Doug Fields <dfields-pg-general@pexicom.com> writes: > > (gdb) where > > #0 0x4028299d in fdatasync () from /lib/libc.so.6 > > #1 0x081049ae in pg_fdatasync (fd=53) at fd.c:233 > > #2 0x08088276 in issue_xlog_fsync () at xlog.c:3367 > >Hmm. That leads to a different line of thought about where the problem >is. Does the problem go away if you turn off fsync, or select a >wal_sync_method other than fdatasync? (See postgresql.conf; not sure >if you need a postmaster restart or just SIGHUP to change these.) Interestingly enough, I was just testing that (the fsync off part). FSYNC=OFF results: I just turned fsync=off and completely restarted the postmaster (always figure that's the safest thing to do). What a difference. My checkpoints are just as long - they still take quite a bit of time to process and do a lot of "blocks out" on the vmstat - but queries now longer block waiting for them to finish. Now, of course, the question becomes, with FSYNC=ON still enabled, which WAL_SYNC_METHOD works best. As you note, fdatasync is the default. Let me play some more... FSYNC=ON results: WAL_SYNC_METHOD settings: FDATASYNC - problem exists (the original problem) OPEN_SYNC - problem persists (same problem; things pile up behind a checkpoint) OPEN_DATASYNC - gives runtime error: FATAL 1: invalid value for option 'WAL_SYNC_METHOD': 'OPEN_DATASYNC' and refuses to start FSYNCA - gives runtime error: FATAL 1: invalid value for option 'WAL_SYNC_METHOD': 'FSYNCA' and refuses to start These last two are obviously not compiled into the backend by default on Debian. Of note, in the fdatasync() man page: Currently (Linux 2.2) fdatasync is equivalent to fsync. Additional thoughts? Thanks, Doug
At 7/8/02 6:28 AM, Sander Steffann wrote: >I know someone who does that too (even after telling him it is NOT Postgre a >LOT of times)... I just don't know where people get that idea! I suspect they get it in part from the logo on the home page: It says "Postgre" in dark letters, and then "SQL" in light letters, and it looks like two separate words. Perhaps a feature of a Web redesign should be that the word be all one color, to emphasize the fact that it's pronounced as a single word. ------------------------------------ Robert L Mathews, Tiger Technologies
> > I was able to set a breakpoint at CreateCheckPoint - gdb never found a
> > MoveOfflineLogs for me to set a breakpoint.
>
>Grumble. It's a static routine, and I imagine its name has been totally
>stripped from the executable.
I can now set the breakpoint there with the unstripped executable.
It seems that the MoveOfflineLogs takes very little time; "fin" returns
almost immediately from it.
However, between the breakpoint at CreateCheckPoint and the one at
MoveOfflineLogs takes quite a bit of time.
See transcript: (starting from the completion of a previous "checkpoint"
command and just about to issue another) My notes are embedded with *** NOTE
gdb) where
#0 CreateCheckPoint (shutdown=0 '\0') at xlog.c:3125
#1 0x08111066 in ProcessUtility (parsetree=0x837b468, dest=Remote,
completionTag=0xbfffdf10 "") at utility.c:778
#2 0x0810ecc5 in pg_exec_query_string (query_string=0x837b2d8 "CHECKPOINT;",
dest=Remote, parse_context=0x834e628) at postgres.c:766
#3 0x0810fd5e in PostgresMain (argc=4, argv=0xbfffe140,
username=0x8329e29 "dfields") at postgres.c:1926
#4 0x080f6d4e in DoBackend (port=0x8329cf8) at postmaster.c:2243
#5 0x080f669f in BackendStartup (port=0x8329cf8) at postmaster.c:1874
#6 0x080f5882 in ServerLoop () at postmaster.c:995
#7 0x080f5391 in PostmasterMain (argc=1, argv=0x8312448) at postmaster.c:771
#8 0x080d4e18 in main (argc=1, argv=0xbfffeac4) at main.c:206
(gdb) c
Continuing.
Breakpoint 1, CreateCheckPoint (shutdown=0 '\0') at xlog.c:2912
2912 in xlog.c
(gdb) c
Continuing.
*** NOTE: it now takes a bit of time to get to this next breakpoint
Breakpoint 2, MoveOfflineLogs (log=14, seg=88, endptr=
{xlogid = 14, xrecoff = 1556353340}) at xlog.c:1604
1604 in xlog.c
(gdb) fin
Run till exit from #0 MoveOfflineLogs (log=14, seg=88, endptr=
{xlogid = 14, xrecoff = 1556353340}) at xlog.c:1604
*** NOTE: This "fin" command returns almost immediately
CreateCheckPoint (shutdown=0 '\0') at xlog.c:3125
3125 in xlog.c
(gdb) c
Continuing.
With the recompiled non-stripped postgres binary, connecting to one of the
blocked threads during a checkpoint:
pexicast_lg=# select * from pg_stat_activity where datname='pexicast_lg';
datid | datname | procpid | usesysid | usename
|
current_query
---------+-------------+---------+----------+---------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
3926366 | pexicast_lg | 5130 | 100 | tomcat | UPDATE
list_entries SET last_mail=now() WHERE list_entry_id IN
(623793,623813,623833,623853,623873,623893,623913,623933,623953,623973,623993,624013,624033,624053,624073,624093,624113,624133,624153,624173,624193,624213,624233,624253,624273,624293,624313,62
(gdb) where
#0 0x4028299d in fdatasync () from /lib/libc.so.6
#1 0x081049ae in pg_fdatasync (fd=53) at fd.c:233
#2 0x08088276 in issue_xlog_fsync () at xlog.c:3367
#3 0x08085206 in XLogWrite (WriteRqst=
{Write = {xlogid = 14, xrecoff = 1170656964}, Flush = {xlogid =
14, xrecoff = 1170656964}}) at xlog.c:1167
#4 0x080854fd in XLogFlush (record={xlogid = 14, xrecoff = 1170656964})
at xlog.c:1277
#5 0x0808335b in RecordTransactionCommit () at xact.c:610
#6 0x0808365e in CommitTransaction () at xact.c:973
#7 0x08083929 in CommitTransactionCommand () at xact.c:1256
#8 0x0810eedf in finish_xact_command () at postgres.c:931
#9 0x0810edee in pg_exec_query_string (
query_string=0x837b1b8 "UPDATE list_entries SET last_mail=now() WHERE
list_entry_id IN
(623793,623813,623833,623853,623873,623893,623913,623933,623953,623973,623993,624013,624033,624053,624073,624093,624113,624133,624153,624"...,
dest=Remote,
parse_context=0x834e508) at postgres.c:871
#10 0x0810fd5e in PostgresMain (argc=4, argv=0xbfffe140,
username=0x83297e9 "tomcat") at postgres.c:1926
#11 0x080f6d4e in DoBackend (port=0x83296b8) at postmaster.c:2243
#12 0x080f669f in BackendStartup (port=0x83296b8) at postmaster.c:1874
#13 0x080f5882 in ServerLoop () at postmaster.c:995
#14 0x080f5391 in PostmasterMain (argc=1, argv=0x8312448) at postmaster.c:771
#15 0x080d4e18 in main (argc=1, argv=0xbfffeac4) at main.c:206
Oddly enough, I can no longer control-C out after doing a "c"ontinue
command. Hm.
Cheers,
Doug
At 02:17 PM 7/8/2002, Tom Lane wrote:
>Doug Fields <dfields-pg-general@pexicom.com> writes:
> > I was able to set a breakpoint at CreateCheckPoint - gdb never found a
> > MoveOfflineLogs for me to set a breakpoint.
>
>Grumble. It's a static routine, and I imagine its name has been totally
>stripped from the executable.
>
> > Also, I can't seem to recompile PostgreSQL because Debian can't find a
> > tclConfig.sh.
>
>Oliver, what's an appropriate way to rebuild PG with debug symbols from
>the Debian package?
>
> regards, tom lane
Doug Fields <dfields-pg-general@pexicom.com> writes:
> It seems that the MoveOfflineLogs takes very little time; "fin" returns
> almost immediately from it.
> However, between the breakpoint at CreateCheckPoint and the one at
> MoveOfflineLogs takes quite a bit of time.
Well, that's not unexpected; checkpoint is going to issue a deal of I/O
and then sync() it. But that should *not* cause blockage of other
backends; at worst they should slow down a bit due to I/O contention.
Can you do some more research on the stack traces of the stuck backends?
Don't assume they are all alike...
regards, tom lane
Doug Fields <dfields-pg-general@pexicom.com> writes:
> (gdb) where
> #0 0x4028299d in fdatasync () from /lib/libc.so.6
> #1 0x081049ae in pg_fdatasync (fd=53) at fd.c:233
> #2 0x08088276 in issue_xlog_fsync () at xlog.c:3367
Hmm. That leads to a different line of thought about where the problem
is. Does the problem go away if you turn off fsync, or select a
wal_sync_method other than fdatasync? (See postgresql.conf; not sure
if you need a postmaster restart or just SIGHUP to change these.)
regards, tom lane
On Mon, 8 Jul 2002, Greg Sabino Mullane wrote: > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > > Nothing. We even have audio files on the website so there is > > no question on how to pronounce it. Some folks just aren't > > happy unless they can change what doesn't need changing is > > all. I guess it's their way of contributing. > > I don't understand some of the animosity I have seen towards > what I consider a positive change. It obviously does need > changing, otherwise so many people would not be discussing it. Funny, that's just about the same argument the pro GPL people use when the licensing discussion comes up. > An audio file on the website is not going to help people who > are reading the word inside text somewhere. The very fact that > such an audio file even needs to exist should be telling us > something. It was created because people were having a hard time > pronouncing it. Saying that people can pronounce it properly > because there is an audio file smacks of circular logic. That's not what was said. Some people can play it over and over and still not be able to pronounce it. Same with my last name and yours for all that matter. But that's no reason to change it, or are you willing to change your last name so all of use can pronounce it? The audio file is there so if someone wants to know how it's really pronounced they can click on it and listen. Tell me, do you pronounce "Linux" the same way Linus does or some other way? Should "Linux" be changed to something that has a more common pronunciation? And yes, I know how Linus pronounces it, I've heard it - someone sent me an mp3 of it. Vince. -- ========================================================================== Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net 56K Nationwide Dialup from $16.00/mo at Pop4 Networking Online Campground Directory http://www.camping-usa.com Online Giftshop Superstore http://www.cloudninegifts.com ==========================================================================
> I don't understand some of the animosity I have seen towards
> what I consider a positive change. It obviously does need
> changing, otherwise so many people would not be discussing it.
Well, it is the same reaction you would have if someone walked up to you
and said "your child has a stupid name. You need to give it another
one". At least you didn't call our child ugly too ;)
I know that "you" is also one of "us", but the point is mostly the
same...
afaict the name has not inhibited our market acceptance. But maybe I'm
giving too much credit to suits' abilities to cope.
PgSQL works for me if I'm wanting an all-acronym acronym, and I've been
known to write and use "Postgres" as a synonym. But historically we
needed to differentiate the name from PostQuel-enabled Postgres. It
could be worse: we could have stuck with Postgres95 (ack, spit).
- Thomas
On Mon, 8 Jul 2002, Greg Sabino Mullane wrote: > Sorry, but the shortcut "Postgres" is alive and kicking. Look anywhere, > even in the mailing lists. The default user is "postgres" not > "postgresql". Another point is that postgres fits in the 8.3 naming > schema Sorry, but you just lost your argument the moment you throw out M$/DOS standards as something to live by ...
On Mon, 8 Jul 2002, Doug Fields wrote: > >Dunno. The CHECKPOINT would probably create a significant number of > >disk write requests, followed by a sync() request. If that could > >monopolize an ext3 filesystem for a long time, perhaps that would > >explain your problem. But I haven't heard similar complaints before. > > > >What do you have shared_buffers set to? > > pexicast_lg=# show shared_buffers; > NOTICE: shared_buffers is 65536 > SHOW VARIABLE > > I believe this is about 512 megs. I have 8 gigs RAM on this server and > tried 65536 and before it 32768. Either one seems to work fine - I didn't > notice any significant performance difference yet between them. Try 1024, and see if it makes a difference. If it still doesn't help, bring up your system with whatever kernel boot parameter you need to to make it think there's only, 256 MB of RAM in the system, and see if the problem goes away. What kind of disk subsystem do you have on this machine? How many MB/sec can you write to the data drives when doing sequential writes? Random writes? Do you have anything else (such as your logfile) on the same drives as the data files? Outside of the disk with the transaction log, do you see any disk activity between checkpoints? Or, just after a checkpoint, do you see almost no disk activity (except to the transaction log)? I'm wondering if perhaps your cached filesystem data blocks are not being written out as they are generated, but are being saved up and only written when a checkpoint occurs (and they are forced out). Since you can cache so many changed blocks, you would have to have a very, very capable disk subsystem to be able to deal with hundreds of megabytes being written out all at once, and in more or less random order. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
On Mon, 8 Jul 2002, Tom Lane wrote: > Well, that's not unexpected; checkpoint is going to issue a deal of I/O > and then sync() it. But that should *not* cause blockage of other > backends; at worst they should slow down a bit due to I/O contention. Well, depending on how the OS schedules writes, one process doing a huge amount of writing might well slow down everything else a lot, unless you've got a really good disk system. But is it possible for a process to commit a transaction while a checkpoint is in progress? That would mean that it's ok for the checkpoint record to be after a bunch of transactions that are not part of the checkpoint, right? cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
Curt Sampson <cjs@cynic.net> writes:
> Well, depending on how the OS schedules writes, one process doing
> a huge amount of writing might well slow down everything else a
> lot, unless you've got a really good disk system.
It's currently looking like an fdatasync applied to a single file
might get blocked behind a global sync in his system. Does that
ring a bell with anyone?
> But is it possible for a process to commit a transaction while a
> checkpoint is in progress?
You bet.
> That would mean that it's ok for the
> checkpoint record to be after a bunch of transactions that are not
> part of the checkpoint, right?
Right. The physical position of the checkpoint record is distinct
from its logical position, if you will. There's a back-pointer in
the checkpoint record showing where the end of WAL was when the
checkpointer started, and any replay has to start there, not at the
physical position of the checkpoint record. Records in between were
emitted while the checkpoint ran, and so their effects might or might
not have been synced to disk by the checkpoint.
regards, tom lane
Doug Fields <dfields-pg-general@pexicom.com> writes:
> FSYNC=OFF results:
> What a difference. My checkpoints are just as long - they still take quite
> a bit of time to process and do a lot of "blocks out" on the vmstat - but
> queries now longer block waiting for them to finish.
This is beginning to look like a kernel problem.
> FSYNCA - gives runtime error: FATAL 1: invalid value for option
> 'WAL_SYNC_METHOD': 'FSYNCA'
> and refuses to start
That should be FSYNC not FSYNCA, I believe.
> Of note, in the fdatasync() man page: Currently (Linux 2.2)
> fdatasync is equivalent to fsync.
Are you really using a 2.2 kernel? I thought that that point had been
fixed in 2.4 kernels. (I see the comment is still there in the man page
on my RH7.2 box, though.)
Now that we've eliminated MoveOfflineLogs as the time-consuming part of
checkpoint, would you make the same check for mdsync()? That's the
routine that actually issues the sync() calls. Again, it would be
useful to know how long it takes, and whether the other backends appear
to be blocked at the times mdsync() is entered and exited.
regards, tom lane
On Mon, 8 Jul 2002, Vince Vielhaber wrote:
> That's not what was said. Some people can play it over and over and
> still not be able to pronounce it. Same with my last name and yours
> for all that matter. But that's no reason to change it, or are you
> willing to change your last name so all of use can pronounce it?
Woah! I gather you are a techie, and don't care much about marketing, hmm?
Personally, if I were trying to market myself to a broader audience
than I have now, I would almost certainly change my name to something
less confusing, more memorable and easier to pronounce. And many others
agree with me, judging by the number of celebrities and wanna-be
celebrities that have done so. Not to mention products that have been
rebranded every decade or three to change with the changing tastes of
the target market.
A lot of the things that market-oriented folks want to do (such as
changing a name) may seem stupid and useless to you, as a techie, but
that doesn't mean that they don't work. It just means that they don't
work on people like you, who comprise a very small market. (I'm in that
market too, by the way; I just recognise that most people do not make
decisions the way I do about what to purchase and/or use.)
I don't know if you were one of the ones complaining about postgres not
being so popular (as, say, compared to MySQL), but if you want it to be
more popular, trying to stop the folks interesting in marketing it from
marketing it is not the way to help things.
> Tell me, do you pronounce "Linux" the same way Linus does or some
> other way? Should "Linux" be changed to something that has a more
> common pronunciation? And yes, I know how Linus pronounces it, I've
> heard it - someone sent me an mp3 of it.
This is a quite different situation. Linux is almost never misspelled,
or broken up into two incorrect pieces, and Linux is dead easy to
pronounce in English, even if it's not the same ("correct") Finnish
pronounciation. The incorrect pronounciation of "Linux" is no worse
a problem than your undoubtedly incorrect pronounciation of "Nissan,"
"Toyota", "Tokyo", and many other Japanese words (or the Japanese
pronounciation of many popular foreign names).
cjs
--
Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org
Don't you know, in this new Dark Age, we're all light. --XTC
On Mon, 8 Jul 2002, Thomas Lockhart wrote: > > I don't understand some of the animosity I have seen towards > > what I consider a positive change. It obviously does need > > changing, otherwise so many people would not be discussing it. > > Well, it is the same reaction you would have if someone walked up to you > and said "your child has a stupid name. You need to give it another > one". At least you didn't call our child ugly too ;) Another way of putting it ... alot of ppl name their child 'Samantha', but how many refer to them as 'Sam' in everyday conversation? One of their 'formal, on paper' name, the other is their common name *shrug*
Curt Sampson wrote: > On Mon, 8 Jul 2002, Vince Vielhaber wrote: > > > That's not what was said. Some people can play it over and over and > > still not be able to pronounce it. Same with my last name and yours > > for all that matter. But that's no reason to change it, or are you > > willing to change your last name so all of use can pronounce it? > > Woah! I gather you are a techie, and don't care much about marketing, hmm? > > Personally, if I were trying to market myself to a broader audience > than I have now, I would almost certainly change my name to something > less confusing, more memorable and easier to pronounce. And many others > agree with me, judging by the number of celebrities and wanna-be > celebrities that have done so. Not to mention products that have been > rebranded every decade or three to change with the changing tastes of > the target market. > > A lot of the things that market-oriented folks want to do (such as > changing a name) may seem stupid and useless to you, as a techie, but > that doesn't mean that they don't work. It just means that they don't > work on people like you, who comprise a very small market. (I'm in that > market too, by the way; I just recognise that most people do not make > decisions the way I do about what to purchase and/or use.) > > I don't know if you were one of the ones complaining about postgres not > being so popular (as, say, compared to MySQL), but if you want it to be > more popular, trying to stop the folks interesting in marketing it from > marketing it is not the way to help things. I totally agree. The name has proven to be hard to pronounce, and that is bad for marketing, period. Is marketing important enough to change the name? That is the question. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Tue, 9 Jul 2002, Bruce Momjian wrote: > I totally agree. The name has proven to be hard to pronounce, and that > is bad for marketing, period. Is marketing important enough to change > the name? That is the question. No, period. For starters, ppl are confusing 'word of mouth' with marketing, which they aren't the same ... Marketing *is* the 8 or so books on the shelves and at Amazon for PostgreSQL, and the ones to follow ... Marketing is the thousands of t-shirts and mugs and CDs that have gone out over the past 4+ years ... Marketing is the countless articles/reviews that ppl have written that talk about PostgreSQL ... Marketing is the awards we have won over the years ... Marketing is proliferation of the newsgroups over the 'Net over the past 4+ years ... Marketing is the countless companies out there that offer PostgreSQL services, support *and* training ... If ppl want to be lazy and call it Postgres, so be it ... as I mentioned before, its like calling Samantha, Sam ... but the formal name itself is, and will remain, PostgreSQL ... its what *alot* of us have been marketing for years now, period.
Marc G. Fournier wrote: > On Tue, 9 Jul 2002, Bruce Momjian wrote: > > > I totally agree. The name has proven to be hard to pronounce, and that > > is bad for marketing, period. Is marketing important enough to change > > the name? That is the question. > > No, period. > > For starters, ppl are confusing 'word of mouth' with marketing, which they > aren't the same ... > > Marketing *is* the 8 or so books on the shelves and at Amazon for > PostgreSQL, and the ones to follow ... So you think, even marketing-wise that PostgreSQL is better. At a minimum, we should add "also called 'postgres'" to all our introductory documentation. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Tue, 9 Jul 2002, Bruce Momjian wrote: > Marc G. Fournier wrote: > > On Tue, 9 Jul 2002, Bruce Momjian wrote: > > > > > I totally agree. The name has proven to be hard to pronounce, and that > > > is bad for marketing, period. Is marketing important enough to change > > > the name? That is the question. > > > > No, period. > > > > For starters, ppl are confusing 'word of mouth' with marketing, which they > > aren't the same ... > > > > Marketing *is* the 8 or so books on the shelves and at Amazon for > > PostgreSQL, and the ones to follow ... > > So you think, even marketing-wise that PostgreSQL is better. At a > minimum, we should add "also called 'postgres'" to all our introductory > documentation. I do agree with this ... in fact, it shoudl go as far as saying something like: PostgreSQL (aka PgSQL aka Postgres aka Pg) ... ppl use all the various forms ... Note that the lists themselves act as their own marketing ... pgsql-*@postgresql.org ... and all the search engines have postgresql.org in them, and, I'm sorry, but someones lame argument about 'whether to search for postgres or postgresql' ... like, come on ... if you have any doubt, just search for postgres, it *is* a sub-string of the formal name ...
At 04:12 PM 7/8/2002, Tom Lane wrote: >This is beginning to look like a kernel problem. Sigh. The hardest ones to fix. It's very nice however that the fsync=off fixes the immediate problem, inasmuch as I can run with that for the moment until such a time as we figure out the "real" problem. >That should be FSYNC not FSYNCA, I believe. OK I'll try that. > > Of note, in the fdatasync() man page: Currently (Linux 2.2) > > fdatasync is equivalent to fsync. > >Are you really using a 2.2 kernel? I thought that that point had been >fixed in 2.4 kernels. (I see the comment is still there in the man page >on my RH7.2 box, though.) No, I'm using 2.4.18 on Debian/Woody, but as you say, the man page still refers to 2.2. >Now that we've eliminated MoveOfflineLogs as the time-consuming part of >checkpoint, would you make the same check for mdsync()? That's the >routine that actually issues the sync() calls. Again, it would be >useful to know how long it takes, and whether the other backends appear >to be blocked at the times mdsync() is entered and exited. Will do. Thanks, Doug
I hate to play devil's advocate, but isn't the reason it is called PostgreSQL now due to marketing? As I understand it, the name PostgreSQL came about because the development team wanted to convey the fact that Postgres95 was now an SQL based DBMS. From a technical standpoint there's no reason it couldn't be called Postgres2002 or some such nonsense, but it might be more cryptic as to what purpose it serves. That's marketing, pure and simple. Robert Treat On Tue, 2002-07-09 at 13:09, Marc G. Fournier wrote: > On Tue, 9 Jul 2002, Bruce Momjian wrote: > > > I totally agree. The name has proven to be hard to pronounce, and that > > is bad for marketing, period. Is marketing important enough to change > > the name? That is the question. > > No, period. > > For starters, ppl are confusing 'word of mouth' with marketing, which they > aren't the same ... > > Marketing *is* the 8 or so books on the shelves and at Amazon for > PostgreSQL, and the ones to follow ... > > Marketing is the thousands of t-shirts and mugs and CDs that have gone out > over the past 4+ years ... > > Marketing is the countless articles/reviews that ppl have written that > talk about PostgreSQL ... > > Marketing is the awards we have won over the years ... > > Marketing is proliferation of the newsgroups over the 'Net over the past > 4+ years ... > > Marketing is the countless companies out there that offer PostgreSQL > services, support *and* training ... > > If ppl want to be lazy and call it Postgres, so be it ... as I mentioned > before, its like calling Samantha, Sam ... but the formal name itself is, > and will remain, PostgreSQL ... its what *alot* of us have been marketing > for years now, period. >
At 7/9/02 11:09 AM, Marc G. Fournier wrote: >I'm sorry, but someones lame argument about 'whether to >search for postgres or postgresql'... like, come on ... if you have any >doubt, just search for postgres, it *is* a sub-string of the formal name Hmmm? Most search engines do not treat substrings in this manner. Try the following two searches on Google: statistics gathering postgresql damond walker statistics gathering postgres damond walker The latter search does not return any matches, while the first does. (Obviously, the reverse is also true -- searching for "postgresql" won't show you matches of pages that only mention "postgres".) In general, if you do two searches on "postgres" and "postgresql", the results are quite different, usually in the opposite direction from which you suggested: you get more hits for "postgresql". This effect makes searches difficult for PostgreSQL users, and makes the software appear less popular to people doing searches on only one term, especially "postgres" (or, heaven forbid, "postgre"). The name MySQL, for example, does not have this drawback, although I suppose some people might search for "my sql" and have a similar problem. Perhaps you consider this a trivial effect, but it is real and I have often found results for obscure problems with one search and not the other, so I don't think my original comment was "lame" :-) ------------------------------------ Robert L Mathews, Tiger Technologies
Wow, cool, I didn't realize that ... but, interesting enough, if you do just: postgresql statistics gathering or postgres statistics gathering the two first links contain *only* postgresql, so its gotta be pulling a substring in from *somewhere* ... On Tue, 9 Jul 2002, Robert L Mathews wrote: > At 7/9/02 11:09 AM, Marc G. Fournier wrote: > > >I'm sorry, but someones lame argument about 'whether to > >search for postgres or postgresql'... like, come on ... if you have any > >doubt, just search for postgres, it *is* a sub-string of the formal name > > Hmmm? Most search engines do not treat substrings in this manner. Try the > following two searches on Google: > > statistics gathering postgresql damond walker > statistics gathering postgres damond walker > > The latter search does not return any matches, while the first does. > (Obviously, the reverse is also true -- searching for "postgresql" won't > show you matches of pages that only mention "postgres".) In general, if > you do two searches on "postgres" and "postgresql", the results are quite > different, usually in the opposite direction from which you suggested: > you get more hits for "postgresql". > > This effect makes searches difficult for PostgreSQL users, and makes the > software appear less popular to people doing searches on only one term, > especially "postgres" (or, heaven forbid, "postgre"). The name MySQL, for > example, does not have this drawback, although I suppose some people > might search for "my sql" and have a similar problem. > > Perhaps you consider this a trivial effect, but it is real and I have > often found results for obscure problems with one search and not the > other, so I don't think my original comment was "lame" :-) > > ------------------------------------ > Robert L Mathews, Tiger Technologies > > > ---------------------------(end of broadcast)--------------------------- > TIP 3: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@postgresql.org so that your > message can get through to the mailing list cleanly >
> >I'm sorry, but someones lame argument about 'whether to > >search for postgres or postgresql'... like, come on ... if you have any > >doubt, just search for postgres, it *is* a sub-string of the formal name > > Hmmm? Most search engines do not treat substrings in this manner. Try the > following two searches on Google: > > statistics gathering postgresql damond walker > statistics gathering postgres damond walker > > The latter search does not return any matches, while the first does. > (Obviously, the reverse is also true -- searching for "postgresql" won't > show you matches of pages that only mention "postgres".) In general, if > you do two searches on "postgres" and "postgresql", the results are quite > different, usually in the opposite direction from which you suggested: > you get more hits for "postgresql". > > This effect makes searches difficult for PostgreSQL users, and makes the > software appear less popular to people doing searches on only one term, > especially "postgres" (or, heaven forbid, "postgre"). The name MySQL, for > example, does not have this drawback, although I suppose some people > might search for "my sql" and have a similar problem. Actually they don't... if you do a google search on "my sql query" the first thing google reports back is: Did you mean: mysql query -philip
On Tue, 9 Jul 2002, Bruce Momjian wrote: >Curt Sampson wrote: >> Personally, if I were trying to market myself to a broader audience >> than I have now, I would almost certainly change my name to something >> less confusing, more memorable and easier to pronounce. And many others >> agree with me, judging by the number of celebrities and wanna-be >> celebrities that have done so. Not to mention products that have been >> rebranded every decade or three to change with the changing tastes of >> the target market. While I think I agree with Curt, I don't thing that it is an *urgent* problem that is plaguing Postgres[QL]. Funny thing on /. today with regards to OpenBeOS interviews: " The answer to all the 'is there room in the market?' questions was answered in a way: 'We are an OSS project. Marketing is not our job.' " Although I do wish that Postgres had better market presence. It is certainly my DB of choice, but many people I have to work with/for will first ask 'do you support MySQL?'. Now we've *had* to support MySQL, but I believe that this is only because of its popularity and market presence. I'd love to tell them all 'no, we support Postgres[QL] which offers a whole lot more'. Cheers, Chris -- Christopher Murtagh Webmaster / Sysadmin Web Communications Group McGill University Montreal, Quebec Canada Tel.: (514) 398-3122 Fax: (514) 398-2017
On Tue, 9 Jul 2002, Marc G. Fournier wrote: > PostgreSQL (aka PgSQL aka Postgres aka Pg) ... ppl use all the various > forms ... Having to say that is a very, very good reason to just change the name back to "postgres" *everywhere*, and stick with that. > Note that the lists themselves act as their own marketing ... > pgsql-*@postgresql.org ... and all the search engines have > postgresql.org in them.... Is there a difficulty in using postgres.org instead? > , and, I'm sorry, but someones lame argument about 'whether to search > for postgres or postgresql' ... like, come on ... if you have any > doubt, just search for postgres, it *is* a sub-string of the formal > name I did search. A search for "postgres" turns up only one quarter of the hits, does not turn up the advertisement for postgresql documentation, and the second link is to a page called "University POSTGRES 4.2". The advertisement thing worries me particularly; it means that advertisers have to advertise on more keywords in order to achieve reasonable coverage. Now, having done a web search on "pgsql", I can now see your difficulty with the name change; you are the president of a company called "PostgreSQL Inc." Why didn't you just say this from the beginnning? Nobody here is trying to make life difficult for those promoting postgres, and if you really are marketing the "PostgreSQL" name hard, maybe we shouldn't change it. But I'm not seeing a lot of evidence of that marketing, unfortunately (or perhaps I would have heard of your company before this). cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
On Wed, 10 Jul 2002, Curt Sampson wrote: > On Tue, 9 Jul 2002, Marc G. Fournier wrote: > > > PostgreSQL (aka PgSQL aka Postgres aka Pg) ... ppl use all the various > > forms ... > > Having to say that is a very, very good reason to just change the name > back to "postgres" *everywhere*, and stick with that. It will *not* happen, so you may as well just drop that part of the thread. > > Note that the lists themselves act as their own marketing ... > > pgsql-*@postgresql.org ... and all the search engines have > > postgresql.org in them.... > > Is there a difficulty in using postgres.org instead? Is there a difficulty in accepting that it will not change? > > , and, I'm sorry, but someones lame argument about 'whether to search > > for postgres or postgresql' ... like, come on ... if you have any > > doubt, just search for postgres, it *is* a sub-string of the formal > > name > > I did search. A search for "postgres" turns up only one quarter of the > hits, does not turn up the advertisement for postgresql documentation, > and the second link is to a page called "University POSTGRES 4.2". Guess you should learn to type 'postgresql' then, eh? > Now, having done a web search on "pgsql", I can now see your difficulty > with the name change; you are the president of a company called > "PostgreSQL Inc." Why didn't you just say this from the beginnning? Because it was irrelevant? And still is ...
On Wed, 10 Jul 2002, Marc G. Fournier wrote: > It will *not* happen, so you may as well just drop that part of the > thread. And so who died and appointed you king? Sorry, but PostgreSQL is not your product, much as you might like to think so. And I find it rather offensive that you should pretend it is. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
On Tue, 9 Jul 2002, Marc G. Fournier wrote: > the two first links contain *only* postgresql, so its gotta be pulling a > substring in from *somewhere* ... I don't believe that Google will match "postgres" with "postgresql" at all. Most likely one of the pages that referred to that page had "postgres" in it somewhere. (It's not well known that google will find pages where only a page that links to it contains a term in your search list.) cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
Can you please stop this madness? Or maybe this thread was started in order to kill PostgreSQL in the first place? If you keep going on like this someone will finish by noticing and that will only hurt PostgreSQL and everything that has been achieved up til now. WE DO NOT NEED A NAME CHANGE. The name is just fine as it is. Tony Grant -- RedHat Linux on Sony Vaio C1XD/S http://www.animaproductions.com/linux2.html Macromedia UltraDev with PostgreSQL http://www.animaproductions.com/ultra.html
On 10 Jul 2002, tony wrote: > Can you please stop this madness? Or maybe this thread was started in > order to kill PostgreSQL in the first place? Hardly. > If you keep going on like this someone will finish by noticing and that > will only hurt PostgreSQL and everything that has been achieved up til > now. So then stop your temper tantrum, stop flaming people who are just considering the idea, and contribute something useful to the discussion. Yelling in all caps that we don't need to change the name, while providing no support for your argument, is certainly not going to help the heat-light ratio here. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
On Wed, 2002-07-10 at 10:24, Curt Sampson wrote: > So then stop your temper tantrum, stop flaming people who are just > considering the idea, and contribute something useful to the discussion. > Yelling in all caps that we don't need to change the name, while > providing no support for your argument, is certainly not going to help > the heat-light ratio here. If you don't want people to take PostgreSQL seriously then just put a big red sign on the first page of the web site. "The free rdbms that often changes it name - come back often to see how we are called" seems like a good choice to me... The whole discussion is a stupid waste of time because there is _nothing_ wrong with the name as it is now. It is only causing pain and confusion amongst users, newcomers and developers. It is only serving to hurt PostgreSQL and how it is perceived. Please stop! Or just tell us that this sabotage is an effort to make us lose all that we have invested in PostgreSQL? How many times a year will I have to go and tell my clients that they are now using "xxxxx" as their SQL database server? Will they understand? It is hard enough as it is selling open source in some parts. We just don't need your help in making it worse. Changing product names without reason does not make sense. In yesterdays post I said that Coke and Coca-Cola are the same drink. Postgres and PostgreSQL are the same database for most sane people that I have met in this community. I just want the nutters to shut up and let us get on with something more serious. This is not a flame - it is common sense. Tony -- RedHat Linux on Sony Vaio C1XD/S http://www.animaproductions.com/linux2.html Macromedia UltraDev with PostgreSQL http://www.animaproductions.com/ultra.html
On 10 Jul 2002, tony wrote: > If you don't want people to take PostgreSQL seriously then just put a > big red sign on the first page of the web site. > > "The free rdbms that often changes it name - come back often to see how > we are called" seems like a good choice to me... That's a definite downside, yes. On the other hand, we essentially already have three names at the moment: PostgreSQL, Postgres, and PgSQL. Is it better to perpetuate that confusion forever, or to make a concerted effort to fix it and fix the problem for good? > The whole discussion is a stupid waste of time because there is > _nothing_ wrong with the name as it is now. "No, you're a stupid moron because you can't see the problem." Is this the response you're looking for? If not, I suggest you put away your flamethrower. > It is only causing pain and > confusion amongst users, newcomers and developers. It is only serving to > hurt PostgreSQL and how it is perceived. Funny, that's the exact complaint some people have about the current name problem. > I just want the nutters to shut up and let us get on with something more > serious. "Then be quiet." > This is not a flame - it is common sense. To call me and others on this list a "nutter" and to call our ideas "stupid" and a "waste of time" is not a flame? Come on now. Anyway, feel free to flame me back and get the last word in, if you like. I shan't be replying to you again unless you come back without the insults. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
On Wednesday 10 Jul 2002 5:45 am, Curt Sampson wrote: > On Tue, 9 Jul 2002, Marc G. Fournier wrote: > > the two first links contain *only* postgresql, so its gotta be pulling a > > substring in from *somewhere* ... > > I don't believe that Google will match "postgres" with "postgresql" > at all. Well, why don't we ask them to? The "did you mean..." message is obviously something they can configure. Does anyone on the list have contacts at Google? Failing that, a polite request from the core developers will probably yield results and good public relations for all concerned. PS - I use "PostgreSQL", "Postgres", "pgsql" and even "PG" depending on how tired with typing I am. There is nothing wrong with the official name (imho) but if we are going to use alternatives it'd be nice if any searches would match on it. - Richard Huxton
On Wed, 10 Jul 2002, Richard Huxton wrote: > The "did you mean..." message is obviously something they can configure. I don't see why it's "obviously" something they can configure, if you mean it's easy to do and they're prepared to do it for us. I suspect that they're just recording historical data on searches, or looking for similar words that matched the found documents and a lot more, or something like that. However, a request to have them always suggest PostgreSQL when someone types in postgres might be worthwhile, if we decide we really want to stick with PostgreSQL for ever and ever. > PS - I use "PostgreSQL", "Postgres", "pgsql" and even "PG" depending on how > tired with typing I am. There is nothing wrong with the official name (imho) > but if we are going to use alternatives it'd be nice if any searches would > match on it. I don't think it's likely we'd ever achieve that on a broad basis. There are just too many search engines out there.... cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
> -----Original Message----- > From: Marc G. Fournier [mailto:scrappy@hub.org] > Subject: Re: [GENERAL] I am being interviewed by OReilly > > On Wed, 10 Jul 2002, Curt Sampson wrote: > > Having to say that is a very, very good reason to just > change the name > > back to "postgres" *everywhere*, and stick with that. > > It will *not* happen, so you may as well just drop that part of the > thread. ... > Guess you should learn to type 'postgresql' then, eh? which will not happen so we may as well drop the other half of the thread. I call it postgres, always have, always will. Because I'm initiated I know the special case for accessing the right website. End of story for me. Sam
Hi there,
On 10 Jul 2002, tony wrote:
> Can you please stop this madness? Or maybe this thread was started in
> order to kill PostgreSQL in the first place?
>
> If you keep going on like this someone will finish by noticing and that
> will only hurt PostgreSQL and everything that has been achieved up til
> now.
>
> WE DO NOT NEED A NAME CHANGE. The name is just fine as it is.
Although I think that a name change is a complicated action to
take, I don't know why all the resistance to this discussion. Until now,
the only argument I saw from people that are against the change is that
PostgreSQL's name was/is being recognised with very hard work. However, I
really don't see where a simple name change would avoid people from
recognising and deploying postgres/postgres sql/postgresql.
For me it seems to be a positive and critical move. It not only
would solve the issue, but would show people that the project is really up
and kicking, not only in the software department, but also as related to
market acceptance. And, good or not, PostgreSQL is now a world class
product that cannot be considered a "hobby", both for the developers and
users. Just to mention MySQL, it's much more known and deployed. Why?
Software? Don't think so. It's name is simple, it's logo is great, and it
has a nice homepage and users advocating. Success is not just product
quality.
And, as a fan of PostgreSQL, of course I do advocate it. But it's
not easy at all tell people about postgresql. When you'll tell them
it's features, advantages, people often break your speaking with a string
attached: "Post what?".
[]'s
Ricardo.
On Wednesday 10 July 2002 12:37 am, Curt Sampson wrote: > On Wed, 10 Jul 2002, Marc G. Fournier wrote: > > It will *not* happen, so you may as well just drop that part of the > > thread. > And so who died and appointed you king? > Sorry, but PostgreSQL is not your product, much as you might like to > think so. And I find it rather offensive that you should pretend it is. Curt, you do realize that Marc Fournier helped start this whole thing (taking over Postgres95 from its two developers and founding PostgreSQL), is a founding member of the PostgreSQL steering committee (core), administers/runs/pays for the postgresql.org website, coordinates and performs the actual work of the release, and many other things that are necessary. His contributions to the project entitle him to have far more say than you have. If you doubt that fact, you need to read the archives for awhile to get a sense of how this project is organized. If the steering committee (the core six) decide against something, then that something _does_not_happen_. End of story. This is not a democracy. It is an oligarchy. Marc is one of the six oligarchs, so _Deal_with_it_. Bruce, another of the core six, has to an extent agreed with some of the difficulty of the current name. But how have the rest weighed in? Up until the last portions of this thread I might have agreed with you to an extent. But after I weighed the difficulty of actually pulling off a name change, I am dead set against it. It's too much effort for too little gain. Now, if you want to pay for the bandwidth of a 'postgres.org', want to set up a full CVS repository, want to administer a popular server, and want to evangelize enough developers to gather a critical mass to fork a 'postgres' project, then go ahead. The name is fine as it is. This is because: 1.) It is well known by that name; 2.) Books are already written using that name (there are no books about 'Postgres'; 3.) It is descriptive; 4.) It has history; 5.) "If it ain't broke, don't fix it" -- and the name ain't broke. The whole project should not have to deal with all the ramifications of a name change just for a few people's convenience, laziness, and stubborness. -- Lamar Owen WGCR Internet Radio 1 Peter 4:11
On Wed, Jul 10, 2002 at 10:55:45AM -0300, suga@netbsd.com.br wrote:
> users. Just to mention MySQL, it's much more known and deployed. Why?
> Software? Don't think so. It's name is simple, it's logo is great, and it
> has a nice homepage and users advocating. Success is not just product
> quality.
I don't actually care at all what the-RDBMS-we-use-now is called, and
I'm loathe to add fuel to the fire, but I doubt very much that
MySQL's success is due to its name and logo. I suspect it has
something of a "first mover" advantage: when people wanted to set up
quick-and-dirty SQL databases, MySQL was "free enough" and "good
enough" for the purposes. So some people (notably Slashdot) used it,
and that caused others to use it, and now there are a lot of people
using it.
PostgreSQL was originally something of an academic project, and it
was seriously buggy and slow in its early days. It's gained a lot of
ground, but in the absence of MySQL-killer features, people don't
move. It turns out that PostgreSQL _has_ some killer features; when
people need them, they do move. But for most purposes (like web
pages), MySQL is still good enough. (It always seemed like a toy to
me, but then I always needed a system that would not lose data no
matter what. I gather it's improved, so don't start a flamefest
about that -- I just don't care.)
PostgreSQL's real marketing problem is that it's too good to be
needed by the MySQL crowd (or they switch), and it's not a big
commercial package for the we-need-a-big-commercial-package crowd.
It's far from obvious that a name change would help in the latter
case, which is the real market potential. I wish I know what _would_
help that, though, in the absence of (sigh) Great Bridge.
A
--
----
Andrew Sullivan 87 Mowat Avenue
Liberty RMS Toronto, Ontario Canada
<andrew@libertyrms.info> M6K 3E3
+1 416 646 3304 x110
Replication would help that. Our company would gladly use Postgres instead of MS SQL Server if postgres could do replication. On Wed, 10 Jul 2002, Andrew Sullivan wrote: > PostgreSQL's real marketing problem is that it's too good to be > needed by the MySQL crowd (or they switch), and it's not a big > commercial package for the we-need-a-big-commercial-package crowd. > It's far from obvious that a name change would help in the latter > case, which is the real market potential. I wish I know what _would_ > help that, though, in the absence of (sigh) Great Bridge.
On Wed, Jul 10, 2002 at 10:33:39AM -0700, Ben wrote:
> Replication would help that. Our company would gladly use Postgres instead
> of MS SQL Server if postgres could do replication.
The replication offered by PostgreSQL, Inc. is pretty good, and
certainly good enough for most people's cases. It's commercial. It
certainly works. That it doesn't do hot failover is a gotcha,
though.
A
--
----
Andrew Sullivan 87 Mowat Avenue
Liberty RMS Toronto, Ontario Canada
<andrew@libertyrms.info> M6K 3E3
+1 416 646 3304 x110
PostgreSQL, Inc has a commercial replication available, and I understand there is a fair amount of activity going on with the PgReplication project as well ... On Wed, 10 Jul 2002, Ben wrote: > Replication would help that. Our company would gladly use Postgres instead > of MS SQL Server if postgres could do replication. > > On Wed, 10 Jul 2002, Andrew Sullivan wrote: > > > PostgreSQL's real marketing problem is that it's too good to be > > needed by the MySQL crowd (or they switch), and it's not a big > > commercial package for the we-need-a-big-commercial-package crowd. > > It's far from obvious that a name change would help in the latter > > case, which is the real market potential. I wish I know what _would_ > > help that, though, in the absence of (sigh) Great Bridge. > > > ---------------------------(end of broadcast)--------------------------- > TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org >
I just looked at the PgSQL Inc page, on the projects page there is a link to
www.erserver.com should this be considered current ?
How can one see more details about the commercial replication [maybe even
including a price ?] with out buying it ?
Jeff.
-----Original Message-----
From: pgsql-general-owner@postgresql.org
[mailto:pgsql-general-owner@postgresql.org]On Behalf Of Andrew Sullivan
Sent: Wednesday, July 10, 2002 2:46 PM
To: postgres list
Subject: Re: [GENERAL] I am being interviewed by OReilly
On Wed, Jul 10, 2002 at 10:33:39AM -0700, Ben wrote:
> Replication would help that. Our company would gladly use Postgres instead
> of MS SQL Server if postgres could do replication.
The replication offered by PostgreSQL, Inc. is pretty good, and
certainly good enough for most people's cases. It's commercial. It
certainly works. That it doesn't do hot failover is a gotcha,
though.
A
--
----
Andrew Sullivan 87 Mowat Avenue
Liberty RMS Toronto, Ontario Canada
<andrew@libertyrms.info> M6K 3E3
+1 416 646 3304 x110
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
On Fri, Aug 09, 2002 at 02:53:48PM -0300, Jeff MacDonald wrote:
> I just looked at the PgSQL Inc page, on the projects page there is
> a link to www.erserver.com should this be considered current ?
>
> How can one see more details about the commercial replication
> [maybe even including a price ?] with out buying it ?
I don't know the answer to these, sorry; only the PostgreSQL, Inc.
guys do. I _can_ tell you that the product works similarly to the
rserv code in contrib/, only it has been enhanced, and is
considerably faster. It _is_ somewhat raw: there are no whizz-bang
administration features, and recovering from a slave failure requires
some effort; nevertheless, it works, and works well even under heavy
load.
A
--
----
Andrew Sullivan 87 Mowat Avenue
Liberty RMS Toronto, Ontario Canada
<andrew@libertyrms.info> M6K 3E3
+1 416 646 3304 x110
Oh, yes, I understand PgReplication is comming nicely, and that will be great then - but it isn't production quality today. Unfortunately, I was unaware of PostgreSQL, Inc's commercial replication. Of course, now that all our stored proceedures are written for MS SQL Server, we'll need a Very Good Reason to leave. :( On Wed, 10 Jul 2002, Marc G. Fournier wrote: > > PostgreSQL, Inc has a commercial replication available, and I understand > there is a fair amount of activity going on with the PgReplication project > as well ... > > On Wed, 10 Jul 2002, Ben wrote: > > > Replication would help that. Our company would gladly use Postgres instead > > of MS SQL Server if postgres could do replication. > > > > On Wed, 10 Jul 2002, Andrew Sullivan wrote: > > > > > PostgreSQL's real marketing problem is that it's too good to be > > > needed by the MySQL crowd (or they switch), and it's not a big > > > commercial package for the we-need-a-big-commercial-package crowd. > > > It's far from obvious that a name change would help in the latter > > > case, which is the real market potential. I wish I know what _would_ > > > help that, though, in the absence of (sigh) Great Bridge. > > > > > > ---------------------------(end of broadcast)--------------------------- > > TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org > > >
> It will *not* happen, so you may as well just drop that part of the > thread. Hmm. But anyway, why not register the domain postgres.org in addition to postgresql.org, no matter if the name will change or not. Let it point to the same server, forward to postgresql.org or whatever. It might make a few new people look the right place :-) -- Kaare Rasmussen --Linux, spil,-- Tlf: 3816 2582 Kaki Data tshirts, merchandize Fax: 3816 2501 Howitzvej 75 Åben 12.00-18.00 Email: kar@webline.dk 2000 Frederiksberg Lørdag 11.00-17.00 Web: www.suse.dk
On Wed, 10 Jul 2002, Kaare Rasmussen wrote: > > It will *not* happen, so you may as well just drop that part of the > > thread. > > Hmm. But anyway, why not register the domain postgres.org in addition to > postgresql.org, no matter if the name will change or not. Already is ... and it *was* supposed to be pointing at postgresql.org ... getting that fixed now :(
call us at the office or send to sales@pgsql.com ... have to get our web designer to go through that site with a fine-toothed comb ... On Fri, 9 Aug 2002, Jeff MacDonald wrote: > I just looked at the PgSQL Inc page, on the projects page there is a link to > www.erserver.com should this be considered current ? > > How can one see more details about the commercial replication [maybe even > including a price ?] with out buying it ? > > Jeff. > > -----Original Message----- > From: pgsql-general-owner@postgresql.org > [mailto:pgsql-general-owner@postgresql.org]On Behalf Of Andrew Sullivan > Sent: Wednesday, July 10, 2002 2:46 PM > To: postgres list > Subject: Re: [GENERAL] I am being interviewed by OReilly > > > On Wed, Jul 10, 2002 at 10:33:39AM -0700, Ben wrote: > > Replication would help that. Our company would gladly use Postgres instead > > of MS SQL Server if postgres could do replication. > > The replication offered by PostgreSQL, Inc. is pretty good, and > certainly good enough for most people's cases. It's commercial. It > certainly works. That it doesn't do hot failover is a gotcha, > though. > > A > > -- > ---- > Andrew Sullivan 87 Mowat Avenue > Liberty RMS Toronto, Ontario Canada > <andrew@libertyrms.info> M6K 3E3 > +1 416 646 3304 x110 > > > ---------------------------(end of broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org > > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster >
I'd like to be able to get a hash (e.g. md5) of a text datatype inside the database. It looks like this is not included in the default installation. Has anyone rolled one of these? "Weak" checksum algorithms (e.g. crypt) would work, too. Mike
On Wed, Jul 10, 2002 at 03:38:49PM -0400, Michael Adler wrote: > I'd like to be able to get a hash (e.g. md5) of a text datatype inside the > database. It looks like this is not included in the default installation. Try contrib/pgcrypto Cheers, Neil -- Neil Conway <neilconway@rogers.com> PGP Key ID: DB3C29FC
Thanks for the heads-up Neil. I had trouble loading this library. Any suggestions on debuging this? user@localhost:~/work/pgsql/contrib/pgcrypto$ make installcheck make -C ../../src/test/regress pg_regress make[1]: Entering directory `/home/adler/work/pgsql/src/test/regress' make[1]: `pg_regress' is up to date. make[1]: Leaving directory `/home/adler/work/pgsql/src/test/regress' ../../src/test/regress/pg_regress init md5 sha1 hmac-md5 hmac-sha1 blowfish rijndael crypt-des crypt-md5 crypt-blowfish crypt-xdes (using postmaster on Unix socket, default port) ============== dropping database "regression" ============== DROP DATABASE ============== creating database "regression" ============== CREATE DATABASE ALTER DATABASE ============== dropping regression test user accounts ============== ERROR: DROP GROUP: group "regressgroup1" does not exist ============== installing PL/pgSQL ============== ============== running regression test queries ============== test init ... ERROR: Load of file /usr/local/pgsql/lib/pgcrypto.so failed: /usr/local/pgsql/lib/pgcrypto.so: undefined symbol: px_find_dige st ERROR: Load of file /usr/local/pgsql/lib/pgcrypto.so failed: /usr/local/pgsql/lib/pgcrypto.so: undefined symbol: px_find_digest On Wed, 10 Jul 2002, Neil Conway wrote: > Date: Wed, 10 Jul 2002 15:46:31 -0400 > From: Neil Conway <nconway@klamath.dyndns.org> > To: Michael Adler <adler@glimpser.org> > Cc: postgres list <pgsql-general@postgresql.org> > Subject: Re: [GENERAL] hash(text) function inside db? > > On Wed, Jul 10, 2002 at 03:38:49PM -0400, Michael Adler wrote: > > I'd like to be able to get a hash (e.g. md5) of a text datatype inside the > > database. It looks like this is not included in the default installation. > > Try contrib/pgcrypto > > Cheers, > > Neil > > -- > Neil Conway <neilconway@rogers.com> > PGP Key ID: DB3C29FC > Mike
"Marc G. Fournier" <scrappy@hub.org> writes:
>> Hmm. But anyway, why not register the domain postgres.org in addition to
>> postgresql.org, no matter if the name will change or not.
> Already is ... and it *was* supposed to be pointing at postgresql.org ...
> getting that fixed now :(
Since we also have postgres.com, don't forget to make that point to the
right place too.
GB did get a couple of things done anyway ;-)
regards, tom lane
Lamar Owen wrote: > If you doubt that fact, you need to read the archives for awhile to get a > sense of how this project is organized. If the steering committee (the core > six) decide against something, then that something _does_not_happen_. End of > story. This is not a democracy. It is an oligarchy. Marc is one of the six > oligarchs, so _Deal_with_it_. Bruce, another of the core six, has to an > extent agreed with some of the difficulty of the current name. But how have > the rest weighed in? Up until the last portions of this thread I might have > agreed with you to an extent. But after I weighed the difficulty of actually > pulling off a name change, I am dead set against it. It's too much effort > for too little gain. I don't think you can just "shut down" a discussion about a name change. Some good things are coming out of is, such as adding "also called 'postgres'" to some of our documentation, and properly mapping postgres.org/com to postgresql.org. I think there is room for an "also called postgres" push among our users and for marketing. Oracle is changing the name of their server all the time to position it for marketing so having a secondary name doesn't hurt. Our _official_ name is PostgreSQL. (I personally voted for 'tigres' at the time we chose PostgreSQL.) -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Wed, 10 Jul 2002, Lamar Owen wrote: > Curt, you do realize that Marc Fournier.... Sure. You do realize that I've been using postgres on and off since before Postgres95 existed, right? Regardless, if people think it's not worthwhile to change the name, I have no problem with that. It's the, "Just shut up because I don't want to listen to you attitude" that I have a problem with. Not to mention the, "It's wrong, but I don't have to give any reasons why, because I'm just right and you're not" attitude. You know, I just bought this up as an idea to be batted around. And I, for one, am still not sure whether changing the name is a good idea or not. But I find really disappointing the number of people here who a) object to an idea for no good reason that they can explain, beyond "I don't like it," and b) are flaming those who are discussing this idea rather than dismissing it. > This is not a democracy. It is an oligarchy. Marc is one of the six > oligarchs, so _Deal_with_it_. So, basically, "if you don't like the product, get lost." If this is really the case, we can certainly drop any discussion of the name, because we have much, much bigger marketing problems. > 1.) It is well known by that name; Not to mention by one or two other names. > 4.) It has history; Though less history than other names. > 5.) "If it ain't broke, don't fix it" -- and the name ain't broke. Well, apparently some people don't agree with you. You can listen to them or you can tell them to get lost. Your choice. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
On Thu, 11 Jul 2002, Curt Sampson wrote: > On Wed, 10 Jul 2002, Lamar Owen wrote: > > > Curt, you do realize that Marc Fournier.... > > Sure. You do realize that I've been using postgres on and off since > before Postgres95 existed, right? Well you certainly don't act like it - and before you go snapping at me for saying that, go back and look at your previous posts and ask yourself why you would ask some of the questions or make some of the statements you made if you really had been around that long. Vince. -- ========================================================================== Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net 56K Nationwide Dialup from $16.00/mo at Pop4 Networking Online Campground Directory http://www.camping-usa.com Online Giftshop Superstore http://www.cloudninegifts.com ==========================================================================
Hello all: I could use some recommendations for what kind of hardware to throw at my database system. It will be postgres 7.2 running under RedHat, most likely on Intel hardware. The application is not very large. The data is probably no more than 5 gig, no query returns more than a few thousand rows, and the joins, though frequent, are of at best medium complexity, generally involving no more than 4-5 tables. The application is front-ended by PHP running under Apache on Linux. The database activity is probably moderately biased toward reading rather than writing. Individual users tend to pull up one fairly complex record at a time (special education forms) and edit it page by page. The biggest concern is not really query speed, or even raw transaction speed, but scalability under multiple users. What do I need, hardware-wise, to try to maximize the number of simultaneous open connections? I'm considering a fairly middle-of-the-road Dell PowerEdge, 2 gig RAM, 1.2GHz Xeon, and possibly a RAID 5 array. Is this overkill? How much, if at all, will PG benefit from additional processors? Thanks for any thoughts, steve
On Thu, 11 Jul 2002, Steve Lane wrote: > I'm considering a fairly middle-of-the-road Dell PowerEdge, 2 gig RAM, > 1.2GHz Xeon, and possibly a RAID 5 array. Since you said you've got a fairly small amount of data (5GB), definitely go with mirroring rather than RAID. It's much faster. cjs -- Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're all light. --XTC
I need to get rid of all rows that have dups in the columns tpa,pun,grn,claim ... i.e. 1--- 001 001 001 00-000001 John Doe 2--- 001 001 001 00-000001 Jane Doe 3--- 001 002 001 00-000001 John Doe 1 and 2 would be dups, 1 and 3 are diff records, 2 and 3 are diff records. I tried this as a test: select count(claimid), tpa, pun, grn, claim FROM claim_import GROUP BY tpa, pun, grn, claim HAVING count(claimid) > 1; 26 rows returned. then select distinct on (tpa,pun,grn,claim) count(claimid), tpa, pun, grn, claim FROM claim_import GROUP BY tpa, pun, grn, claim HAVING count(claimid) > 1; and still had 26 rows returned. not sure how that can happen. Anyway, then I tried CREATE UNIQUE INDEX tmpidx ON claim (tpa,pun,grn,claim); that of course failed, stating dups existed. Any help would be appreciated. Thanks, Jeremy
Curt Sampson wrote: > > On Wed, 10 Jul 2002, Lamar Owen wrote: > > > Curt, you do realize that Marc Fournier.... > > Sure. You do realize that I've been using postgres on and off since > before Postgres95 existed, right? Have been there since v4.2 (last official Berkeley release, PostQUEL version) and survived the dark era (Postgres95). That by itself is no argument, nor does it give someones word more weight. PostgreSQL is the name of this software for many years, books are printed using that name, press articles refer to it, marketing is based on it. If we like it or not, it is IMHO not an option to change it just to sound better or because people with spelling difficulties don't get it right. Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). Some people here in the US think I'm a girl, most pronounce my first name Dshaen or worse many fail on my last name somewhere around OUUEEEE... cough, cough. Do you think I consider changing something? No way, not for people who speak one or less languages only. PostgreSQL is PostgreSQL, and that's it. Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #================================================== JanWieck@Yahoo.com #
> (I personally voted for 'tigres' at the time we chose PostgreSQL.) ?? What about "digress" ?? :-) -- Kaare Rasmussen --Linux, spil,-- Tlf: 3816 2582 Kaki Data tshirts, merchandize Fax: 3816 2501 Howitzvej 75 Åben 12.00-18.00 Email: kar@webline.dk 2000 Frederiksberg Lørdag 11.00-17.00 Web: www.suse.dk
Hello,
Ahhh that is not true. RAID 5 is much, much faster than a mirror (RAID 1). I believe you are thinking about mirrored + striped which is RAID10 (also know as 1+0). That is faster than RAID 5 but requires 4 disks where RAID 5 requires 3.
J
Curt Sampson wrote:
Ahhh that is not true. RAID 5 is much, much faster than a mirror (RAID 1). I believe you are thinking about mirrored + striped which is RAID10 (also know as 1+0). That is faster than RAID 5 but requires 4 disks where RAID 5 requires 3.
J
Curt Sampson wrote:
On Thu, 11 Jul 2002, Steve Lane wrote:I'm considering a fairly middle-of-the-road Dell PowerEdge, 2 gig RAM, 1.2GHz Xeon, and possibly a RAID 5 array.Since you said you've got a fairly small amount of data (5GB), definitely go with mirroring rather than RAID. It's much faster. cjs
Jeremy Cowgar wrote:
> I need to get rid of all rows that have dups in the columns
> tpa,pun,grn,claim ... i.e.
>
> 1--- 001 001 001 00-000001 John Doe
> 2--- 001 001 001 00-000001 Jane Doe
> 3--- 001 002 001 00-000001 John Doe
>
> 1 and 2 would be dups, 1 and 3 are diff records, 2 and 3 are diff
> records.
>
> I tried this as a test:
>
> select count(claimid), tpa, pun, grn, claim FROM claim_import GROUP BY
> tpa, pun, grn, claim HAVING count(claimid) > 1;
> 26 rows returned.
>
> then
>
> select distinct on (tpa,pun,grn,claim) count(claimid), tpa, pun, grn,
> claim FROM claim_import GROUP BY tpa, pun, grn, claim HAVING
> count(claimid) > 1;
It's not obvious to me what your key(s) is (all 3 columns?), but this is
a place where self-joins are useful. Assuming a table like:
create table stuff (
id int, -- primary table key
value int, -- unique data key
...);
You should be able to find the dups with something like:
select b.id
from stuff a, stuff b
where a.value = b.value
and a.id < b.id;
Given that, then use it to get:
delete from stuff
where id in (select b.id from stuff a, stuff b where ...);
Be careful and experiment with the select until you're 110% sure you
like what you see. :-) Adapt this approach to your real table and you
should be set.
HTH,
Kevin
Marc G. Fournier: > I can never figure this out ... what is so difficult about 'Postgres-Q-L'? http://www.postgresql.org/ "Ever wonder how PostgreSQL is really pronounced?" ;-)
On Thu, 11 Jul 2002, Jan Wieck wrote: > Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). Is that how its pronounced?? :) I've known you for how long now and never had a clue how to pronounce your last name ... first is/was easy, last I never even tried ...
Marc G. Fournier wrote: > On Thu, 11 Jul 2002, Jan Wieck wrote: > > > Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). > > Is that how its pronounced?? :) I've known you for how long now and never > had a clue how to pronounce your last name ... first is/was easy, last I > never even tried ... And we did think Jan was a girl for many months until somehow he mentioned something that gave us a clue he wasn't. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Thu, 11 Jul 2002, Bruce Momjian wrote: > And we did think Jan was a girl for many months until somehow he > mentioned something that gave us a clue he wasn't. Through the years I've had many friends that I've never seen and several that I had no idea as to their gender. Isn't e-mail great. No preconceived ideas when we start communicating. (Heck I get to be called a @#$%^& based on merit. :-) There has been a Jan, Lynn, Sandy and a couple others I can't remember but the biggest shock was the phone call I got from some one I'd been communicating with for several years and she said "Hi Rod, this is Charlie". I hadn't had a clue Charlie was a woman. As the cartoon says "On the internet no one knows you're a dog." Rod -- "Open Source Software - Sometimes you get more than you paid for..."
Roderick A. Anderson wrote: > On Thu, 11 Jul 2002, Bruce Momjian wrote: > > > And we did think Jan was a girl for many months until somehow he > > mentioned something that gave us a clue he wasn't. > > Through the years I've had many friends that I've never seen and several > that I had no idea as to their gender. Isn't e-mail great. No > preconceived ideas when we start communicating. (Heck I get to be called I do speak to women differently so it was weird not knowing about Jan. If you have seen the SNL skit about Pat, well, its like that. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Okay, going *way* off subject.... why on earth do you speak to women differently? On Thu, 11 Jul 2002, Bruce Momjian wrote: > Roderick A. Anderson wrote: > > On Thu, 11 Jul 2002, Bruce Momjian wrote: > > > > > And we did think Jan was a girl for many months until somehow he > > > mentioned something that gave us a clue he wasn't. > > > > Through the years I've had many friends that I've never seen and several > > that I had no idea as to their gender. Isn't e-mail great. No > > preconceived ideas when we start communicating. (Heck I get to be called > > I do speak to women differently so it was weird not knowing about Jan. > If you have seen the SNL skit about Pat, well, its like that. > > -- > Bruce Momjian | http://candle.pha.pa.us > pgman@candle.pha.pa.us | (610) 853-3000 > + If your life is a hard drive, | 830 Blythe Avenue > + Christ can be your backup. | Drexel Hill, Pennsylvania 19026 > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster >
Ben wrote: > Okay, going *way* off subject.... why on earth do you speak to women > differently? You don't make sarcastic comments to women, for one thing. I speak less harshly, I guess. I am married, so dating isn't the issue. I used to think "treat them the same" but in reality I think most women prefer that you didn't. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Bruce Momjian wrote: > Marc G. Fournier wrote: > >>On Thu, 11 Jul 2002, Jan Wieck wrote: >> >> >>>Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). >> >>Is that how its pronounced?? :) I've known you for how long now and never >>had a clue how to pronounce your last name ... first is/was easy, last I >>never even tried ... > > > And we did think Jan was a girl for many months until somehow he > mentioned something that gave us a clue he wasn't. > There are plenty of hard to spell/pronounce names on the lists - and I don't want to be the one to start *that* thread - but I have to ask: Mom-jee-arn ? Mom-jy-ann ? Mom-jeen ? Mom-zhon ? Cheers, Kurt (Male).
Bruce Momjian wrote: > > Ben wrote: > > Okay, going *way* off subject.... why on earth do you speak to women > > differently? > > You don't make sarcastic comments to women, for one thing. I speak less > harshly, I guess. I am married, so dating isn't the issue. > > I used to think "treat them the same" but in reality I think most women > prefer that you didn't. Amen. Mike Mascari mascarm@mascari.com
On Wednesday 10 July 2002 09:26 pm, Bruce Momjian wrote: > Lamar Owen wrote: > > If you doubt that fact, you need to read the archives for awhile to get a > > sense of how this project is organized. If the steering committee (the > > core six) decide against something, then that something > I don't think you can just "shut down" a discussion about a name change. Well, you're right. *I* can't shut anything down (except my server...). But my reply was targeted to the 'who died and made you king' ad hominem. > I think there is room for an "also called postgres" push among our users > and for marketing. Oracle is changing the name of their server all the > time to position it for marketing so having a secondary name doesn't > hurt. Our _official_ name is PostgreSQL. Is that like the Artist Formerly Known as Prince? Now abbreviated to 'Artist'? Why beat around the bush about our name? This is confusion. > (I personally voted for 'tigres' at the time we chose PostgreSQL.) As in the river? Does that mean we consider ourselves to be the new seat of power in the world? Hey, maybe I need to change my name to Tiglathpileser or something. Hey, we could then have Sargon as our engine, Pul as our system of scripting, Shalmaneser as the trademark for 'readers don't have to wait for writers', Sennacherib as the patented system for stuffing data that goes beyond the size of a tuple..... As long as a project from Iraq called 'Babylon' lead by a royal Nebuchadnezzar doesn't come around.... Time to retreat to Nineveh.... -- Lamar Owen WGCR Internet Radio 1 Peter 4:11
On Wednesday 10 July 2002 10:19 pm, Curt Sampson wrote: > You know, I just bought this up as an idea to be batted around. And I, > for one, am still not sure whether changing the name is a good idea or > not. A name change will be traumatic for a number of reasons. First, every distributor of the progam will have to change its name. Apache is doing this very thing now, and it has got things in a state of confusion -- the apache webserver is now packaged in a a tarball called 'httpd' -- which, IMNSHO, is arrogant. Next, name changes must be effectively communicated to people. If we are having a hard time communicating our current name, how much harder of a time will we have with communicating the fact that 'we've changed our name because we've given up on communicating our name to people'? Apply Occam's Razor, please. Non sunt multiplicanda entia praeter necessitatem. It is up to the ones who want to change the name to come up with a compelling reason to change it -- it's not the responsibility of those who support the current name to defend the current name in order for it to be kept. My opinion is simply that the current name is fine. The question becomes 'is it worth the work to change the name for the minority's convenience?' I believe the logical answer is 'No. While there are seemingly good reasons for this move, none compel this drastic of an action.' -- Lamar Owen WGCR Internet Radio 1 Peter 4:11
Kurt at iadvance wrote: > Bruce Momjian wrote: > > Marc G. Fournier wrote: > > > >>On Thu, 11 Jul 2002, Jan Wieck wrote: > >> > >> > >>>Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). > >> > >>Is that how its pronounced?? :) I've known you for how long now and never > >>had a clue how to pronounce your last name ... first is/was easy, last I > >>never even tried ... > > > > > > And we did think Jan was a girl for many months until somehow he > > mentioned something that gave us a clue he wasn't. > > > > There are plenty of hard to spell/pronounce names on the lists - and I > don't want to be the one to start *that* thread - but I have to ask: > > Mom-jee-arn ? > Mom-jy-ann ? > Mom-jeen ? > Mom-zhon ? Now, that is a good question. It is MOM-jin. Actually, though the Armenian way to say it is MOOM-ji-an, but I don't say it that way. Maybe I need an MP3 file for that. Vince, can you hook me up? ;-) -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Lamar Owen dijo: > On Wednesday 10 July 2002 09:26 pm, Bruce Momjian wrote: > > (I personally voted for 'tigres' at the time we chose PostgreSQL.) > > As in the river? I don't think so. The river is actually called "Tigris" (that's the spanish version, and the version that generates more results in Google, althought I admit that the english transliteration may be different). Tigres is the spanish plural for "tiger", and of course it keeps the play on "-gres" while suggesting some kind of power (and speed?). But I think Bruce should be actually better informed than me as to what was the origin of the word suggestion in the first place. -- Alvaro Herrera (<alvherre[a]atentus.com>) www.google.com: interfaz de linea de comando para la web.
Kurt at iadvance wrote: > Bruce Momjian wrote: > > Marc G. Fournier wrote: > > > >>On Thu, 11 Jul 2002, Jan Wieck wrote: > >> > >> > >>>Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). > >> > >>Is that how its pronounced?? :) I've known you for how long now and never > >>had a clue how to pronounce your last name ... first is/was easy, last I > >>never even tried ... > > > > > > And we did think Jan was a girl for many months until somehow he > > mentioned something that gave us a clue he wasn't. > > > > There are plenty of hard to spell/pronounce names on the lists - and I > don't want to be the one to start *that* thread - but I have to ask: > > Mom-jee-arn ? > Mom-jy-ann ? > Mom-jeen ? > Mom-zhon ? Sorry for the off-topic, but my wife just corrected the proper Armenian pronuciation of my name. It should be MOHM-juh-yan. I rarely correct anyone who pronounces my name. I give them credit for just trying. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Thu, 11 Jul 2002, Bruce Momjian wrote: > Kurt at iadvance wrote: > > Bruce Momjian wrote: > > > Marc G. Fournier wrote: > > > > > >>On Thu, 11 Jul 2002, Jan Wieck wrote: > > >> > > >> > > >>>Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). > > >> > > >>Is that how its pronounced?? :) I've known you for how long now and never > > >>had a clue how to pronounce your last name ... first is/was easy, last I > > >>never even tried ... > > > > > > > > > And we did think Jan was a girl for many months until somehow he > > > mentioned something that gave us a clue he wasn't. > > > > > > > There are plenty of hard to spell/pronounce names on the lists - and I > > don't want to be the one to start *that* thread - but I have to ask: > > > > Mom-jee-arn ? > > Mom-jy-ann ? > > Mom-jeen ? > > Mom-zhon ? > > Now, that is a good question. It is MOM-jin. Actually, though the > Armenian way to say it is MOOM-ji-an, but I don't say it that way. > > Maybe I need an MP3 file for that. Vince, can you hook me up? ;-) No problem, the broadcaster in question is my biz partner. I'll get with him next week. Vince. -- ========================================================================== Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net 56K Nationwide Dialup from $16.00/mo at Pop4 Networking Online Campground Directory http://www.camping-usa.com Online Giftshop Superstore http://www.cloudninegifts.com ==========================================================================
Alvaro Herrera wrote: > Lamar Owen dijo: > > > On Wednesday 10 July 2002 09:26 pm, Bruce Momjian wrote: > > > > (I personally voted for 'tigres' at the time we chose PostgreSQL.) > > > > As in the river? > > I don't think so. The river is actually called "Tigris" (that's the > spanish version, and the version that generates more results in Google, > althought I admit that the english transliteration may be different). > > Tigres is the spanish plural for "tiger", and of course it keeps the > play on "-gres" while suggesting some kind of power (and speed?). But I > think Bruce should be actually better informed than me as to what was > the origin of the word suggestion in the first place. Tigres (female tiger) was suggested because it is the only common *gres word we could think of. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Lamar Owen wrote: > On Wednesday 10 July 2002 09:26 pm, Bruce Momjian wrote: > > Lamar Owen wrote: > > > If you doubt that fact, you need to read the archives for awhile to get a > > > sense of how this project is organized. If the steering committee (the > > > core six) decide against something, then that something > > > I don't think you can just "shut down" a discussion about a name change. > > Well, you're right. *I* can't shut anything down (except my server...). But > my reply was targeted to the 'who died and made you king' ad hominem. And my reply was to those who wanted to shut down the discussion. Lamar, you weren't one of them. The discussion addressed a serious issue, specifically that though PostgreSQL looks great on paper, with the Postgre* and the SQL, it is hard to pronounce. That was the issue, and people are free to make suggestions on how to address it. I think the only discussions we really shut down are those that insult people or are blatantly disrespectful. I think anything else is open for discussion. > > I think there is room for an "also called postgres" push among our users > > and for marketing. Oracle is changing the name of their server all the > > time to position it for marketing so having a secondary name doesn't > > hurt. Our _official_ name is PostgreSQL. > > Is that like the Artist Formerly Known as Prince? Now abbreviated to > 'Artist'? Why beat around the bush about our name? This is confusion. Sure, just call us Database. Maybe Red Hat Database is a trend. ;-) Actually, Prince did that because his record company had rights to his name for publishing for X years. The contract finally ran out and I think he is back to Prince. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
On Thursday 11 July 2002 11:29 pm, Bruce Momjian wrote: > Lamar Owen wrote: > > Well, you're right. *I* can't shut anything down (except my server...). > > But my reply was targeted to the 'who died and made you king' ad hominem. > And my reply was to those who wanted to shut down the discussion. > Lamar, you weren't one of them. As I was in the cc list, it was my message replied to, and my Asperger's kicked in again, I misunderstood. > The discussion addressed a serious issue, specifically that though > PostgreSQL looks great on paper, with the Postgre* and the SQL, it is > hard to pronounce. You know, I think I brought the issue up once myself. But I further realized that it really isn't hard to pronounce. > Sure, just call us Database. Maybe Red Hat Database is a trend. ;-) We might get confused with dBASE. Hmmm. pBASE? (nah.) -- Lamar Owen WGCR Internet Radio 1 Peter 4:11
...
> I think there is room for an "also called postgres" push among our users
> and for marketing. Oracle is changing the name of their server all the
> time to position it for marketing so having a secondary name doesn't
> hurt. Our _official_ name is PostgreSQL.
We have had the following in our documentation for years (I remember
writing it ;)
4. Terminology and Notation
The terms "PostgreSQL" and "Postgres" will be used interchangeably to
refer to the software that accompanies this documentation.
...
I fairly recently went through and scrubbed the docs, which at the time
were (if I remember correctly) roughly 60/40 split between the two forms
of name. I think it is not helpful to try to emphasize the "short form",
but we have always acknowledged that it exists and folks have always
used both forms. It just isn't that big a deal, and it just isn't that
confusing.
- Thomas
> -----Original Message----- > From: Lamar Owen [mailto:lamar.owen@wgcr.org] > Sent: 12 July 2002 04:38 > To: Bruce Momjian > Cc: pgsql-general@postgresql.org > Subject: Re: [GENERAL] I am being interviewed by OReilly > > You know, I think I brought the issue up once myself. But I > further realized > that it really isn't hard to pronounce. It isn't hard to pronounce; we know from anecdote that it is hard to know-how to pronnounce. And people can't always pause the world and run off to check if there is a web site that has an mp3 on it. I'm not for changing the name, even though I will call it postgres myself. Some of the discussion here exists because postgres.org and postgres.com did not redirect to the postgresql.org website. Now they do, so I think much reason to CHANGE the name instead of acknowledge the alias has gone. BTW we might want www.postgres.org to issue an http redirect to www.postgresql.org if you want to emphasise the real name, though I prefer it how it is :-) Sam
On Thu, 2002-07-11 at 18:37, Joshua D. Drake wrote: > Hello, > > Ahhh that is not true. RAID 5 is much, much faster than a mirror (RAID > 1). I believe you are thinking about mirrored + striped which is RAID10 > (also know as 1+0). That is faster than RAID 5 but requires 4 disks > where RAID 5 requires 3. I'd be careful with such general statements. RAID5 may be much faster than mirroring when reading, especially as the number of disks increases. But with RAID5 writing is usually slower (a typical one-stripe-size write requires at least 2 blocks to be read and written again, hitting 2 disks 2 twice. Only n-1-stripe-size writes (where n = # of disks in the array) can be done without reading first). And, btw, RAID10 and RAID 1+0 is not necessarily the same. Depending on the vendor, one is a striped set of mirrored disks, the other a mirrored set of striped disks. RAID 0+1 is also sometimes used to mean the same thing (one or the other). cheers -- vbi -- secure email with gpg http://fortytwo.ch/gpg
Вложения
Bruce Momjian wrote: > Ben wrote: > >>Okay, going *way* off subject.... why on earth do you speak to women >>differently? > > > You don't make sarcastic comments to women, for one thing. I speak less > harshly, I guess. I am married, so dating isn't the issue. > > I used to think "treat them the same" but in reality I think most women > prefer that you didn't. > I personally prefer to be treated the same. Otherwise i feel a bit like a child , childs are treated with indulgence. I can't stand the feeling not beeing taken seriously. Regards Tina
>>>>Look at my own name. Jan Wieck (correctly pronounced like Yann Veek). Vaguely remembered, then found on the Web... Nicklaus Wirth, the designer of PASCAL, gave a talk once at which he was asked "How do you pronounce your name?". He replied, "You can call me by name, pronouncing it 'Virt', or call be by value, 'Worth'." -- John W Hall <wweexxsseessssaa@telusplanet.net> Calgary, Alberta, Canada. "Helping People Prosper in the Information Age"
Sam Liddicott wrote: > > > -----Original Message----- > > From: Lamar Owen [mailto:lamar.owen@wgcr.org] > > Sent: 12 July 2002 04:38 > > To: Bruce Momjian > > Cc: pgsql-general@postgresql.org > > Subject: Re: [GENERAL] I am being interviewed by OReilly > > > > You know, I think I brought the issue up once myself. But I > > further realized > > that it really isn't hard to pronounce. > > It isn't hard to pronounce; we know from anecdote that it is hard to > know-how to pronnounce. We could create a shockwave or flash file from the mp3 and embed it into the main and download pages. Also we could embed some sound into the postmaster. On startup it searches for audio output capability (/dev/dsp) and reports "Postgres-Q-L database system ready" on successful startup. A talking donkey^H^H^H^H^Hatabase! If that's not an argument for every WEB developer! Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #================================================== JanWieck@Yahoo.com #
Hi there,
> Also we could embed some sound into the postmaster. On startup it
> searches for audio output capability (/dev/dsp) and reports
> "Postgres-Q-L database system ready" on successful startup.
Hey, great ideia. It was always nice to hear "starting netra
internet server" from that pizza box Sun... :-)
[]'s
Ricardo.
On Fri, 12 Jul 2002, Jan Wieck wrote: > Also we could embed some sound into the postmaster. On startup it > searches for audio output capability (/dev/dsp) and reports > "Postgres-Q-L database system ready" on successful startup. Could you imagine this in a place like Rackspace ... imagine several hundred computers all rebooting after some planned power upgrade, all chirping up as the PgSQL server auto-starts? :)
On Fri, Jul 12, 2002 at 11:25:44PM -0300, Marc G. Fournier wrote: > On Fri, 12 Jul 2002, Jan Wieck wrote: > > > Also we could embed some sound into the postmaster. On startup it > > searches for audio output capability (/dev/dsp) and reports > > "Postgres-Q-L database system ready" on successful startup. > > Could you imagine this in a place like Rackspace ... imagine several > hundred computers all rebooting after some planned power upgrade, all > chirping up as the PgSQL server auto-starts? :) Surely they don't all have sound cards and speakers installed? Seriously though, how it's pronounced doesn't seem very important to me. As long as people know how to spell it, what's the problem? -- Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/ > There are 10 kinds of people in the world, those that can do binary > arithmetic and those that can't.
On Thu, 11 Jul 2002 23:16:05 GMT, John Hall wrote:
> >>>>Look at my own name. Jan Wieck (correctly pronounced like Yann Veek).
>
> Vaguely remembered, then found on the Web...
>
> Nicklaus Wirth, the designer of PASCAL, gave a talk once at which he
> was asked "How do you pronounce your name?". He replied, "You can call
> me by name, pronouncing it 'Virt', or call be by value, 'Worth'."
You mean this:
"In Europe they call me Niklaus Wirth; in the US they call me Nickel's worth.
That's because in Europe they call me by name, and in the US by value!"
--
Peter Haworth pmh@edison.ioppublishing.com
"The familiar dot '.' symbol from Internet addresses
is used in this book to terminate sentences."
-- Carlton Egremont III, /Mr. Bunny's Guide to ActiveX/
"Marc G. Fournier" wrote: > > On Fri, 12 Jul 2002, Jan Wieck wrote: > > > Also we could embed some sound into the postmaster. On startup it > > searches for audio output capability (/dev/dsp) and reports > > "Postgres-Q-L database system ready" on successful startup. > > Could you imagine this in a place like Rackspace ... imagine several > hundred computers all rebooting after some planned power upgrade, all > chirping up as the PgSQL server auto-starts? :) Kinda like a room full of Vax systems with an overloaded network? "Pling, Pling, Pliplipling, Plingidipling ...". Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #================================================== JanWieck@Yahoo.com #