Bad link on techdocs

Started by Jean-Christian Imbeaultalmost 23 years ago11 messagesgeneral
Jump to latest
#1Jean-Christian Imbeault
jc@mega-bucks.co.jp

I'm dying to read Bruce Momjian's notes on optimising postgres but the
link given on techdocs as been invalid for the past few weeks:

http://www.ca.postgresql.org/docs/momjian/hw_performance/

Does anyone know where I can find a copy of this or when the techdocs
link will be fixed?

Thanks,

Jean-Christian Imbeault

#2Guillaume LELARGE
gleu@wanadoo.fr
In reply to: Jean-Christian Imbeault (#1)
Re: Bad link on techdocs

Le Mardi 17 Juin 2003 09:11, Jean-Christian Imbeault a écrit :

I'm dying to read Bruce Momjian's notes on optimising postgres but the
link given on techdocs as been invalid for the past few weeks:

http://www.ca.postgresql.org/docs/momjian/hw_performance/

Does anyone know where I can find a copy of this or when the techdocs
link will be fixed?

Try this link for a PDF version:
http://www2.cz.postgresql.org/docs/momjian/hw_performance.pdf?N=A

Don't know if it is the current version, I just found it with google
(http://www.google.com/search?q=hw_performance+postgres&sourceid=mozilla-search&start=0&start=0)

--
Guillaume <!-- http://absfr.tuxfamily.org/ -->.

#3Bruce Momjian
bruce@momjian.us
In reply to: Guillaume LELARGE (#2)
Re: Bad link on techdocs

Guillaume LELARGE wrote:

Le Mardi 17 Juin 2003 09:11, Jean-Christian Imbeault a ?crit :

I'm dying to read Bruce Momjian's notes on optimising postgres but the
link given on techdocs as been invalid for the past few weeks:

http://www.ca.postgresql.org/docs/momjian/hw_performance/

Does anyone know where I can find a copy of this or when the techdocs
link will be fixed?

Try this link for a PDF version:
http://www2.cz.postgresql.org/docs/momjian/hw_performance.pdf?N=A

Don't know if it is the current version, I just found it with google
(http://www.google.com/search?q=hw_performance+postgres&amp;sourceid=mozilla-search&amp;start=0&amp;start=0)

Actually, all that stuff is also at:

http://candle.pha.pa.us/main/writings/computer.html

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073
#4Chris Gamache
cgg007@yahoo.com
In reply to: Bruce Momjian (#3)
PostgreSQL calibration

What's your thought on creating some type of calibration applet that will
divine proper settings for the PostgreSQL performance related variables by
taking into consideration the speed of the machine, amount of RAM, access time
of the filesystem, and the speed of various queries on a standardized (perhaps
non-standardized?) dataset under different conditions?

Is this possible, do-able, or even wanted? Where would one begin constructing
the tests?

I imagine it would automate the tasks that every PostgreSQL administrator goes
through to calibrate his or her installation with the best possible performance
settings...

CG

__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

#5Bruce Momjian
bruce@momjian.us
In reply to: Chris Gamache (#4)
Re: PostgreSQL calibration

Sure, it would be great if we could do it.

---------------------------------------------------------------------------

Chris Gamache wrote:

What's your thought on creating some type of calibration applet that will
divine proper settings for the PostgreSQL performance related variables by
taking into consideration the speed of the machine, amount of RAM, access time
of the filesystem, and the speed of various queries on a standardized (perhaps
non-standardized?) dataset under different conditions?

Is this possible, do-able, or even wanted? Where would one begin constructing
the tests?

I imagine it would automate the tasks that every PostgreSQL administrator goes
through to calibrate his or her installation with the best possible performance
settings...

CG

__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073
#6Tom Lane
tgl@sss.pgh.pa.us
In reply to: Bruce Momjian (#5)
Re: PostgreSQL calibration

Bruce Momjian <pgman@candle.pha.pa.us> writes:

Sure, it would be great if we could do it.

If the program actually derives reliable numbers, it would be great.
It could easily do more harm than good if it gives bogus results.
I think it will be very hard to get reliable rather than bogus results
:-( ... but feel free to try.

regards, tom lane

#7Matt Clark
matt@ymogen.net
In reply to: Tom Lane (#6)
Re: PostgreSQL calibration

Are there really any performance settings of much interest beyond the
shared and non-shared memory settings? Beyond those the interactions
get so complex that automation is probably impossible anyway, and
certain options like fsync = false should never be 'recommended'.

On the other hand, a way of empirically deriving some 'correct'
optimizer parameters for a given machine would be very nice :-)

Matt

Show quoted text

On Wed, 2003-06-18 at 18:07, Tom Lane wrote:

Bruce Momjian <pgman@candle.pha.pa.us> writes:

Sure, it would be great if we could do it.

If the program actually derives reliable numbers, it would be great.
It could easily do more harm than good if it gives bogus results.
I think it will be very hard to get reliable rather than bogus results
:-( ... but feel free to try.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

http://archives.postgresql.org

#8scott.marlowe
scott.marlowe@ihs.com
In reply to: Matt Clark (#7)
Re: PostgreSQL calibration

It would be nice to have something that could suggest settings for the
different *cost* options.

On 18 Jun 2003, matt wrote:

Show quoted text

Are there really any performance settings of much interest beyond the
shared and non-shared memory settings? Beyond those the interactions
get so complex that automation is probably impossible anyway, and
certain options like fsync = false should never be 'recommended'.

On the other hand, a way of empirically deriving some 'correct'
optimizer parameters for a given machine would be very nice :-)

Matt

On Wed, 2003-06-18 at 18:07, Tom Lane wrote:

Bruce Momjian <pgman@candle.pha.pa.us> writes:

Sure, it would be great if we could do it.

If the program actually derives reliable numbers, it would be great.
It could easily do more harm than good if it gives bogus results.
I think it will be very hard to get reliable rather than bogus results
:-( ... but feel free to try.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

http://archives.postgresql.org

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

#9Matt Clark
matt@ymogen.net
In reply to: scott.marlowe (#8)
Re: PostgreSQL calibration

Precisely. Any hints from the real gurus out there as to how that might
be accomplished (or alternatively, reasons why it's hopeless)?

Show quoted text

On Wed, 2003-06-18 at 18:42, scott.marlowe wrote:

It would be nice to have something that could suggest settings for the
different *cost* options.

On 18 Jun 2003, matt wrote:

Are there really any performance settings of much interest beyond the
shared and non-shared memory settings? Beyond those the interactions
get so complex that automation is probably impossible anyway, and
certain options like fsync = false should never be 'recommended'.

On the other hand, a way of empirically deriving some 'correct'
optimizer parameters for a given machine would be very nice :-)

Matt

On Wed, 2003-06-18 at 18:07, Tom Lane wrote:

Bruce Momjian <pgman@candle.pha.pa.us> writes:

Sure, it would be great if we could do it.

If the program actually derives reliable numbers, it would be great.
It could easily do more harm than good if it gives bogus results.
I think it will be very hard to get reliable rather than bogus results
:-( ... but feel free to try.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

http://archives.postgresql.org

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

---------------------------(end of broadcast)---------------------------
TIP 8: explain analyze is your friend

In reply to: Chris Gamache (#4)
Re: PostgreSQL calibration

On June 18, 2003 12:54 pm, Chris Gamache wrote:

What's your thought on creating some type of calibration applet that
will divine proper settings for the PostgreSQL performance related

Have you seen pg_autotune? It's aging... but might be something to play
with.

http://gborg.postgresql.org/project/pgautotune/projdisplay.php

--
Tim Middleton | Cain Gang Ltd | But the trouble was that my hysterical
fit
x@veX.net | www.Vex.Net | could not go on for ever. --Dost (NFTU)

#11Dani Oderbolz
oderbolz@ecologic.de
In reply to: Chris Gamache (#4)
Re: PostgreSQL calibration

Chris Gamache wrote:

What's your thought on creating some type of calibration applet that will
divine proper settings for the PostgreSQL performance related variables by
taking into consideration the speed of the machine, amount of RAM, access time
of the filesystem, and the speed of various queries on a standardized (perhaps
non-standardized?) dataset under different conditions?

Other RDBMS offer such options.
I know "Oracle Expert", its resultls are not very striking.
But surely, there is a demand for such tools.
Oracle expert considers the following data (among others)
- RAM
- Number of CPUs
- I/O bandwidth
- Current Statments in cache
- You can fed it with Statements your Application generates (can also
be captured)

But I think it is hard to get precise results,
as the program needs a lot of intelligence...

Cheers, Dani