relationship between PvA and PvB on a
row-by-row basis.
Have you considered using cursors?
--
Paul Thomas
+--+---+
| Thomas Micro Systems Limited | Software Solutions for Business |
| Computer Consultants | http
index but a type mis-match (e.g, an int4 field referencing
an int8 field)
Either of these will cause a sequential table scan and poor performance.
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for
ularly
of course ;) People will want to know:
- PostgreSQL version
- hardware configuration (SCSI or IDE? RAID level?)
- table schemas
- queries together with EXPLAIN ANALYZE output
also output from utils like vmstat, t
or 2 years now and have yet to discover any key
sequence which makes any sense. But then I don't do drugs so my perseption
is probably at odds with the origators of Emacs ;)
--
Paul Thomas
+--+-+
| Thomas Micro Systems Lim
her queries
are slightly slower than under Oracle on the same
hardware but nothing like this.
Usual questions:
have you vacuumed the table recently?
what are your postgresql.conf settings?
can you show us explain ANALYZE output rather than just explain output?
On 17/06/2004 17:54 Vitaly Belman wrote:
Is it possible to download the Visual Explain only (link)? I only see
that you can donwload the whole ISO (which I hardly need).
You can get it from CVS and build it yourself.
--
Paul Thomas
On 17/06/2004 12:10 Adam Witney wrote:
Will this run on other platforms? OSX maybe?
It's a Java app so it runs on any any platform with a reasonably modern
Java VM.
--
Paul Thomas
+--+-+
| Thomas Micro Systems Li
04','182','153','6','2004','0')
DESC
OFFSET 0 LIMIT 20;
I expect that pg is having to evaluate your function every time it does a
compare within its sort. Something like
SELECT t1.value1,t1.value2,
getday_total(..) AS
y underestimated for
todays
average hardware configuration (1+GHz, 0.5+GB RAM, fast FSB, fast HDD).
It seems to me better strategy to force that 1% of users to "downgrade"
cfg.
than vice-versa.
regards
ch
This has been discussed many times before. Check the
.4.
Yes, I've seen other benchmarks which also show that.
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for
Business |
| Computer Consultants |
http://www.t
bout what the value should be..)
(b) is determined by the dastardly trick of actually sampling the data in
the table!!! That's what analyze does. It samples your table(s) and uses
the result to feeede into it's descision about when to flip between
sequential and index scans.
Hope thi
want
I will gladly give you the information.
Googling threw up
http://spider.tm/apr2004/cstory2.html
Interesting and possibly relevant quote:
"Benchmarks have shown that in certain conditions the anticipatory
algorithm is almost 10 times faster than what 2.4 kernel supports".
HTH
hrough the client-side operations to see what could be eating up the 13
seconds.
Given that the client and server are on different machines, I'm wondering
the bulk of the 13 seconds is due a network mis-configuration
stado = 'A') or (Estado = 'I')),
PRIMARY KEY(CodBanco)
);
select * from icc_m_banco where codbanco = 1;
select * from icc_m_banco where codbanco = 1::int2;
--
Paul Thomas
+--+-+
| Thomas Micro S
eriences anyone?
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for the Smaller
Business |
| Computer Consultants |
http://www.thomas-micro-sy
rpm disk, whilst it will probably have a lower seek time than a 10K rpm
disk, won't have a proportionately (i.e., 2/3rds) lower seek time.
- likelihood of page to be cached in memory by the kernel
That's effective cache s
his
email) which has still got its rpm binaries. My other machines have all
been upgraded from source.
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for the Smaller
Business |
| Computer C
RAM) or do you want to use moderate
settings?"
Something like this, you get the idea.
ISR reading that 7.4 will use a default of shared_beffers = 1000 if the
machine can support it (most can). This alone should make a big difference
in out
st a thought..
Interestingly, float8 indexes do work OK (float8col = 99). I spend a large
part of yesterday grepping through the sources to try and find out why
this should be so. No luck so far but I'm going to keep on trying
gle-processor Intel/AMD
based hardware.
Selfishness and sillyness aside, I'm sure your tests will of interest to
us all. Thanks in advance
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for
to me. I'd start off
with something like
shared_buffers = 2000
sort_mem = 1024
max_coonections = 100
and see how it performs under normal business loading.
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software So
k (faster maybe but could cause data corruption). Try modifying your
program to have connection.setAutoCommit(false) and do a
connection.commit() after say every 100 inserts.
HTH
--
Paul Thomas
+--+-+
| Thomas Micro S
should be increased for your situation?
HTH
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for the Smaller
Business |
| Computer Consultants |
http://www.thomas-micro-systems-ltd.co.uk
dscan, enable_sort, enable_nestloop,
enable_mergejoin or enable_hashjoin have been set to false. Looking at the
source, thats the only way I can see that such large numbers can be
produced.
HTH
--
Paul Thomas
+--+-+
| Thomas
find that the third query is also a lot faster that the
first query.
HTH
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for the Smaller
Business |
| Computer Consultants |
http://www.th
ooses a seq scan for small tables
as the whole table can often be bought into memory with one IO whereas
reading the index then the table would be 2 IOs.
HTH
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Sof
ttp://www.redhat.com/docs/manuals/database/. I'd welcome your oppinions
on
this product.
Thank you for your comments.
It looks like they just wrote a number of GUI versions of the command line
utilities. From what I can tell, its still a standard postgresql database
beh
using connection pooling, try reducing the maximum number
of connections. This will take some of the stress off the database.
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for the Smaller
Bus
, you really should consider using a connection pool
as it removes the overhead of creating and closing connections.
--
Paul Thomas
+--+-+
| Thomas Micro Systems Limited | Software Solutions for the Smaller
Business |
| Com
29 matches
Mail list logo
