On Thu, 2009-01-15 at 19:37 -0600, Jason Long wrote:
> >
> > > I have not looked into the detail of the explain, and I do see visually
> > > that very different plans are being chosen.
> > >
> >
> > It would help to share these plans with us...
> >
> See EXPLAIN ANALYZE below for three
Hi,
ltree and pg_trgm with UTF8 support are available from CVS HEAD, see
See http://archives.postgresql.org/pgsql-committers/2008-06/msg00356.php
http://archives.postgresql.org/pgsql-committers/2008-11/msg00139.php
Oleg
On Fri, 16 Jan 2009, pepone.onrez wrote:
On Sat, Apr 19, 2008 at 6:10 PM,
I am setting up Postgres for OpenSSL + FIPs.
I am compiling Postgres with OpenSSL FIPS library using the
"-with-openssl" option. The question I have is, just doing that
suffice? Or do I have to modify the postgres source code?
Since I read through the OpenSSL FIPS documentation, it mentions to
t
> I have a dynamically built query that will periodically(2 times a day and
> becoming more frequent) make my server totally unresponsive.
does this query involve more than geqo_threshold (default 12) tables?
If so, this most probably is geqo (genetic query optimizer) kicking
in. Try to fiddle wi
On Thu, 15 Jan 2009 12:56:51 -0800
"Joshua D. Drake" wrote:
> I know many perfectly intelligent people that are better served
> through diagrams, pdf and color than a mailing list. Most of them
> make sure geeks like us, *EAT*.
> Does that mean they are not intelligent or perhaps that there
> ta
On Thu, 2009-01-15 at 18:44 -0600, Jason Long wrote:
> The query that hangs the system is requesting a count(*)
> based on some parameters the users selects.
Can you show an example of the full offending query? How big is the
table?
> And maybe 1 in 20 will not complete.
If you really have nothi
I am having a serious problem with my application and I hope someone can
help me out.
This could not happen at a worse time as a consulting firm is at my
clients to recommend a new financial system and the inventory
system(which I developed) keeps locking up.
I have a dynamically built query t
Tom Lane wrote:
I read it like this:
#0 0x0827441d in MemoryContextAlloc () <-- real
#1 0x08274467 in MemoryContextStrdup ()<-- real
#2 0x0826501c in database_getflatfilename () <-- real
#3 0x0826504e in database_getflatfilename () <-- must be write_database_file
#4 0x08
Alvaro Herrera writes:
> Tom Lane wrote:
>> Hmm. This isn't very trustworthy for lack of debug symbols (what we're
>> probably looking at are the nearest global function names before the
>> actual locations).
> The lack of debug symbols makes this all mere guesses though. The
> backtrace did no
Tom Lane wrote:
> Hmm. This isn't very trustworthy for lack of debug symbols (what we're
> probably looking at are the nearest global function names before the
> actual locations). However, it strongly suggests that something is
> broken in the active memory context, and the most likely explanat
On Thu, Jan 15, 2009 at 3:49 PM, Bjørn T Johansen wrote:
> Just wondering if I need to change the defalt values for autovacuum in
> version 8.3.5?
They're fairly good. A good way to see if it's working for you is to
let autovacuum run for a few days with your server handling a normal
load, and
On Sat, Apr 19, 2008 at 6:10 PM, Oleg Bartunov wrote:
> On Sat, 19 Apr 2008, Tom Lane wrote:
>
>> Craig Ringer writes:
>>>
>>> Tom Lane wrote:
I don't really see the problem. I assume from your reference to pg_trgm
that you're using trigram similarity as the prefilter for potentia
Just wondering if I need to change the defalt values for autovacuum in version
8.3.5?
Regards,
BTJ
--
---
Bjørn T Johansen
b...@havleik.no
-
Justin Pasher writes:
> Program terminated with signal 11, Segmentation fault.
> #0 0x0827441d in MemoryContextAlloc ()
> (gdb) bt
> #0 0x0827441d in MemoryContextAlloc ()
> #1 0x08274467 in MemoryContextStrdup ()
> #2 0x0826501c in database_getflatfilename ()
> #3 0x0826504e in database_getf
On Thu, Jan 15, 2009 at 2:24 PM, Jason Long
wrote:
> Scott Marlowe wrote:
> You got me. I have a set of mirrored raptors. I am not sure the disk i/o
> subsystem is a bottleneck.
> The whole DB is 50 mb with minimal users.
Then you're only ever writing to the db, and 50Meg is teeny tiny.
Even m
Tom Lane wrote:
Having debug symbols would be more useful, but unless the binary is
totally stripped, a backtrace might provide enough info without that.
Try it and see if you get any function names in the trace, or only
numbers.
(BTW, does Debian have anything comparable to Red Hat's debuginfo
Justin Pasher writes:
> I'll let you know when I get a chance to get a core dump from the
> process. I assume I will need a version of Postgres built with debug
> symbols for it to be useful? I'm not seeing one in the standard Debian
> repositories, so I might have to compile from source.
Havi
Tom Lane wrote:
Justin Pasher writes:
Richard Huxton wrote:
Segmentation fault - probably a bug or bad RAM.
It's a relatively new machine, but that's obviously a possibility with
any hardware. I haven't seen any other programs experiencing problems on
the box, but the Postgres
Andreas Wenk wrote:
> Yes thats correct with the IP address range. Maybe I did not understand
> the auth concept yet. I thought, that with METHOD set to md5, a md5
> hashed password is required. The password is submitted with the PHP 5
> pg_connect function - as plain text.
It is specified
Justin Pasher writes:
> Richard Huxton wrote:
>> Segmentation fault - probably a bug or bad RAM.
> It's a relatively new machine, but that's obviously a possibility with
> any hardware. I haven't seen any other programs experiencing problems on
> the box, but the Postgres daemon is the one that
Hi Tom,
Tom Lane schrieb:
Andreas Wenk writes:
In pg_hba.conf we have:
# TYPE DATABASEUSERCIDR-ADDRESS METHOD
# "local" is for Unix domain socket connections only
local all all ident sameuser
# IPv4 local connections:
host
Steven Lembark wrote:
I would like to use PSQLFS(http://www.edlsystems.com/psqlfs/)
to store 100 GB of images in PostgreSQL.
Once they are in there I can deal with them. My main purpose is to use
rsync to get the files into the database.
Is there a better way to load 20,000 plus files reliably
In response to Kirk Strauser :
> On Jan 15, 2009, at 2:39 PM, Bill Moran wrote:
>
> > However, it pgpool can't pool connections if each connection has its
> > own username. Not sure what exactly is causing it not to work for
> > you,
> > but that was the first thing that came to mind.
>
> The
Richard Huxton wrote:
Justin Pasher wrote:
Hello,
I have a server running PostgreSQL 8.1.15-0etch1 (Debian etch) that was
recently put into production. Last week a developer started having a problem
with his psql connection being terminated every couple of minutes when he
was running a query
Kirk Strauser wrote:
...
I understand why pooling within a process itself is a good thing.
However, say I have two users running the same program on different
desktop machines. At present, those applications connect with the
same username/password that's tied to the program and not the actu
Scott Marlowe wrote:
On Thu, Jan 15, 2009 at 1:28 PM, Jason Long
wrote:
A faster server.
Well the sever is plenty fast. It has 2 quad core 1600MHz FSB 3.0 GHz Xeon
5472 CPUs and a very light workload.
A few things.
That doesn't make a fast server. The disk i/o subsystem makes a fas
On Jan 15, 2009, at 1:02 PM, Scott Marlowe wrote:
On Thu, Jan 15, 2009 at 1:56 PM, Jason Long
wrote:
Steve Atkins wrote:
I'm sure none of that other than the last actually applies to you,
but
those are
the expectations you set by using HTML email and then insulting all
the list members w
On Jan 15, 2009, at 2:54 PM, Steve Crawford wrote:
If you know that the application does not change GUC variables then
you will probably benefit greatly by using pgbouncer.
Thanks, Steve! That's just the kind of pointer I can use. I've been
using PostgreSQL for years but I've never really
On Thu, Jan 15, 2009 at 1:26 PM, Kirk Strauser wrote:
> On Jan 15, 2009, at 12:30 PM, Steve Crawford wrote:
>
>> But if your application is designed to work well with pooling, it can
>> provide dramatic performance benefits.
>
> I think that's the problem. As I mentioned at one point, a lot of ou
On Thu, Jan 15, 2009 at 1:56 PM, Jason Long
wrote:
> Steve Atkins wrote:
>>
>> I'm sure none of that other than the last actually applies to you, but
>> those are
>> the expectations you set by using HTML email and then insulting all
>> the list members when someone asks you to stop. That's not th
On Thu, 2009-01-15 at 20:39 +, Grzegorz Jaśkiewicz wrote:
> and we also oppose to answering on top of message, and citing
> everything underneeth.
> Why? Because your words should say what you mean, not show it by its
> look. Hence, plain ascii is enough for us - and should be for every
> intel
Steve Atkins wrote:
On Jan 15, 2009, at 12:32 PM, Jason Long wrote:
I don't mean to be a pain either and I mean no disrespect to anyone
on this list in the following comments.
However, this is about the most anal list ever.
I see so many emails on here about people complaining regarding the
On Jan 15, 2009, at 2:39 PM, Bill Moran wrote:
However, it pgpool can't pool connections if each connection has its
own username. Not sure what exactly is causing it not to work for
you,
but that was the first thing that came to mind.
The usernames are per-app. Zope connections with usern
Kirk Strauser wrote:
On Jan 15, 2009, at 12:30 PM, Steve Crawford wrote:
But if your application is designed to work well with pooling, it can
provide dramatic performance benefits.
I think that's the problem. As I mentioned at one point, a lot of our
applications have connections open for
On Jan 15, 2009, at 12:32 PM, Jason Long wrote:
I don't mean to be a pain either and I mean no disrespect to anyone
on this list in the following comments.
However, this is about the most anal list ever.
I see so many emails on here about people complaining regarding the
proper way to repl
On 15/01/2009 20:44, Scott Marlowe wrote:
> We're a bunch of fuzzy little kittens
> playing with balls of yarn by comparison. :)
Now *there's* an image! :-)
Ray.
--
Raymond O'Donnell, Director of Music, Galway Cathedral, Ireland
r
On 15/01/2009 20:32, Jason Long wrote:
> However, this is about the most anal list ever. I see so many emails on
> here about people complaining regarding the proper way to reply or post
> to the list.
Well, as someone else has just pointed out, it's all about readability
and making your words ea
On Thu, Jan 15, 2009 at 1:28 PM, Jason Long
wrote:
> A faster server.
> Well the sever is plenty fast. It has 2 quad core 1600MHz FSB 3.0 GHz Xeon
> 5472 CPUs and a very light workload.
A few things.
That doesn't make a fast server. The disk i/o subsystem makes a fast
server. And you've menti
On Thu, 2009-01-15 at 14:32 -0600, Jason Long wrote:
> I don't mean to be a pain either and I mean no disrespect to anyone on
> this list in the following comments.
>
> However, this is about the most anal list ever.
You haven't been to the debian list have you? :).
> I see so many emails on h
and we also oppose to answering on top of message, and citing
everything underneeth.
Why? Because your words should say what you mean, not show it by its
look. Hence, plain ascii is enough for us - and should be for every
intelligent human being.
On Thu, Jan 15, 2009 at 8:32 PM, Jason Long
wrote
In response to Kirk Strauser :
> On Jan 15, 2009, at 12:30 PM, Steve Crawford wrote:
>
> > But if your application is designed to work well with pooling, it
> > can provide dramatic performance benefits.
>
> I think that's the problem. As I mentioned at one point, a lot of our
> application
I don't mean to be a pain either and I mean no disrespect to anyone on
this list in the following comments.
However, this is about the most anal list ever.
I see so many emails on here about people complaining regarding the
proper way to reply or post to the list.
I used larger font to point
Alan Hodgson wrote:
On Thursday 15 January 2009, Jason Long
wrote:
*I am attempting to vacuum and reindex my database. It keeps timing
out. See commands and last part of output below. The vacuum or reindex
only takes a short time to complete normally because the database it
less than 50 m
On Jan 15, 2009, at 12:30 PM, Steve Crawford wrote:
But if your application is designed to work well with pooling, it
can provide dramatic performance benefits.
I think that's the problem. As I mentioned at one point, a lot of our
applications have connections open for hours at a time and
On 15/01/2009 20:21, Joshua D. Drake wrote:
> On Thu, 2009-01-15 at 20:13 +, Raymond O'Donnell wrote:
>> I don't mean to be a pain, but could you please avoid HUGE type sizes
>> such as the aboveor better still, avoid using HTML altogether in
>> your emails to this list.
> The answer to th
On Thu, 2009-01-15 at 20:13 +, Raymond O'Donnell wrote:
> On 15/01/2009 20:06, Jason Long wrote:
>
> I am attempting to vacuum...[snip]
>
> I don't mean to be a pain, but could you please avoid HUGE type sizes
> such as the aboveor better still, avoid using HTML altogether in
> your email
On Thursday 15 January 2009, Jason Long
wrote:
> *I am attempting to vacuum and reindex my database. It keeps timing
> out. See commands and last part of output below. The vacuum or reindex
> only takes a short time to complete normally because the database it
> less than 50 mb. I have the qu
On Thu, 2009-01-15 at 14:06 -0600, Jason Long wrote:
> I am attempting to vacuum and reindex my database. It keeps timing
> out. See commands and last part of output below. The vacuum or
> reindex only takes a short time to complete normally because the
> database it less than 50 mb. I have the
On 15/01/2009 20:06, Jason Long wrote:
I am attempting to vacuum...[snip]
I don't mean to be a pain, but could you please avoid HUGE type sizes
such as the aboveor better still, avoid using HTML altogether in
your emails to this list.
It makes it look as if you are not just shouting, but SCR
*I am attempting to vacuum and reindex my database. It keeps timing
out. See commands and last part of output below. The vacuum or reindex
only takes a short time to complete normally because the database it
less than 50 mb. I have the query timeout set to 2 minutes, but I do
not know if th
Alvaro Herrera schrieb:
Andreas Wenk wrote:
Yes thats correct with the IP address range. Maybe I did not understand
the auth concept yet. I thought, that with METHOD set to md5, a md5
hashed password is required. The password is submitted with the PHP 5
pg_connect function - as plain tex
> On Wed, Jan 14, 2009 at 07:36:03PM -0800, mailingli...@net-virtual.com
> wrote:
>> CREATE TABLE listings (
>> trans_id SERIAL,
>> mode CHAR(1),
>> listing_id INT,
>> region_id INT,
>> category INT
>> );
>>
>> "SELECT * FROM listings ORDER BY region_id, category, listing_id,
>> trans_id
>> Justin Pasher wrote:
> Are there any internal Postgres tables I can look at that may shed some
> light on this? Any particular maintenance commands that could be run for
> repair?
Please obtain a backtrace from the core file. If there's no core file,
please set "ulimit -c unlimited" in th
Hi Joshua
Joshua D. Drake schrieb:
On Thu, 2009-01-15 at 18:05 +0100, Andreas Wenk wrote:
postgres=# SELECT rolname,rolpassword from pg_authid;
rolname | rolpassword
- ---+-
postgres |
pgadmin | plaintext
odie | md5passs
On Thu, Jan 15, 2009 at 11:02 AM, Jeremy Kister
wrote:
> I've got two 300GB databases that I'm going to be upgrading from 8.2.4 (32
> bit) to 8.3.5 (64 bit).
>
> The systems are running on Solaris 10u5 64bit with lots of disks in a zfs
> raid10 and have 32GB ram.
>
> I've read lots of docs and Goo
I've got two 300GB databases that I'm going to be upgrading from 8.2.4 (32
bit) to 8.3.5 (64 bit).
The systems are running on Solaris 10u5 64bit with lots of disks in a zfs
raid10 and have 32GB ram.
I've read lots of docs and Google, and found a special flavor of
postgresql.conf that helps t
On Jan 15, 2009, at 10:20 AM, Bill Moran wrote:
I don't believe that's true. My understanding of pgpool is that it
will
reuse an existing connection if it's free, or open a new one if
required.
Gah! It just made it worse!
$ ps auxwww | grep pgpool | grep dbuser | wc -l
30
$ ps aux
Andreas Wenk writes:
> In pg_hba.conf we have:
> # TYPE DATABASEUSERCIDR-ADDRESS METHOD
> # "local" is for Unix domain socket connections only
> local all all ident sameuser
> # IPv4 local connections:
> hostall all
On Thu, 2009-01-15 at 18:05 +0100, Andreas Wenk wrote:
> postgres=# SELECT rolname,rolpassword from pg_authid;
> rolname | rolpassword
> - ---+-
> postgres |
> pgadmin | plaintext
> odie | md5passsorrrd
>
> The user odi
On Thu, Jan 15, 2009 at 11:57:13AM -0500, Guy Rouillier wrote:
> Connections are pooled on the client end, not on the server end. So,
> you'd be able to pool connections on your web server, and should, for
> reasons documented by others. However, since Abby and Barb are using
> different compu
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi everybody,
I posted this allready to the ADMIN list but recieved no reply (what is for
sure ok in a
way ;-) ). So I thought I'll give it a try here. Sorry for any inconvenience.
We are trying to understand an issue concerning the md5 password enc
On Thu, Feb 21, 2008 at 12:01 PM, Tom Lane wrote:
> Pascal Cohen writes:
>> The fact is that works on Linux and win but under Mac I always get the
>> ordering with 'default' C locale (I displayed all the lc_* and all are
>> right set)
>
> Yeah, this has been complained of before, eg here
> http:/
Kirk Strauser wrote:
I understand why pooling within a process itself is a good thing.
However, say I have two users running the same program on different
desktop machines. At present, those applications connect with the same
username/password that's tied to the program and not the actual us
On Jan 15, 2009, at 10:08 AM, Tom Lane wrote:
As an example, any system catalog update has to be broadcast to all
live backends, and they all have to dutifully search their catalog
caches to flush stale entries. That costs the same whether the
backend is being put to use or has been sitti
Kirk Strauser wrote:
[snip]
> I understand why pooling within a process itself is a good thing.
> However, say I have two users running the same program on different
> desktop machines. At present, those applications connect with the
> same username/password that's tied to the program an
On Thu, Jan 15, 2009 at 10:54 AM, Kirk Strauser wrote:
> I have a PostgreSQL 8.3.5 server with max_connections = 400. At this
> moment, I have 223 open connections, including 64 from a bunch of webserver
> processes and about 100 from desktop machines running a particular
> application. The rest
Tom Lane wrote:
> Kirk Strauser writes:
> > I have a PostgreSQL 8.3.5 server with max_connections = 400. At this
> > moment, I have 223 open connections, including 64 from a bunch of
> > webserver processes and about 100 from desktop machines running a
> > particular application. The rest
Kirk Strauser writes:
> I have a PostgreSQL 8.3.5 server with max_connections = 400. At this
> moment, I have 223 open connections, including 64 from a bunch of
> webserver processes and about 100 from desktop machines running a
> particular application. The rest are from various scheduled
I have a PostgreSQL 8.3.5 server with max_connections = 400. At this
moment, I have 223 open connections, including 64 from a bunch of
webserver processes and about 100 from desktop machines running a
particular application. The rest are from various scheduled processes
and other assorted
Hi,
first, many thanks to all for the great work, i'm waiting for 8.4.
I have played with the new possibilities:
test=# select typ, ts, rank() over (partition by typ order by ts desc ) from
foo;
typ | ts | rank
-+---+--
1 | 20
On Thu, Jan 15, 2009 at 4:57 AM, Christian Schröder wrote:
> Merlin Moncure wrote:
>>> them.
>>> I need something like:
>>> select * from myfunc('mytable') as x(like mytable)
>>> or
>>> select * from myfunc('mytable') as x(mytable%TYPE)
>>>
>>> Is there any solution for PostgreSQL 8.2?
>>>
On Wed, Jan 14, 2009 at 03:56:25PM -0500, Mark Styles wrote:
> SELECT COALESCE(mid,0) AS mid, COALESCE(id_group,0) AS id_group
> FROM users
> WHERE username = 'test'
> UNION
> SELECT 0, 0
> WHERE NOT EXISTS (SELECT 1 FROM users WHERE username = 'test');
An alternative using outer joins would be:
On Wed, Jan 14, 2009 at 07:36:03PM -0800, mailingli...@net-virtual.com wrote:
> CREATE TABLE listings (
> trans_id SERIAL,
> mode CHAR(1),
> listing_id INT,
> region_id INT,
> category INT
> );
>
> "SELECT * FROM listings ORDER BY region_id, category, listing_id,
> trans_id"
> [...] wh
Sorry, I should have RTFM(!!!). I found it under 4.2.4 Field selection.
Apparently it works just as I want, but I should have put parenthesis around
the row-name like this:
> select result,(resulting_row).name from verify_record(1234);
name | result
---|
"Test" | "OK"
I also discove
as far as I know, this bit (statement evaluation) wasn't implemented
then. It only got there in 8.4, so you can have even subselects
evaluated.
So it isn't a bug, it just wasn't implemented to work that way back than,
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make c
Hi,
i'm defining a function in plpqsql and would like it to return one varchar
and one row from another table. I have defined it like this (this is only a
test and does not really make sense yet, but it's the principle i'm after):
CREATE OR REPLACE FUNCTION verify_record(IN number_to_verify bigin
Merlin Moncure wrote:
On 1/13/09, Christian Schröder wrote:
Hi list,
I have written a function that returns a setof record. The function has a
table name as a parameter and the resulting records have the same structure
as this table. Is there any easy way to specify this when I call the
fun
Justin Pasher wrote:
> Hello,
>
> I have a server running PostgreSQL 8.1.15-0etch1 (Debian etch) that was
> recently put into production. Last week a developer started having a problem
> with his psql connection being terminated every couple of minutes when he
> was running a query. When I look th
On Thu, Jan 15, 2009 at 9:27 AM, m zyzy wrote:
>
> Thank you Scott and Dave.
> Dave , the url given , not really helpful unless I go through each and every
> details,
The text at the link I gave reads:
The installer crashes on Linux. What can I do?
BitRock InstallBuilder has been designed
On Thu, Jan 15, 2009 at 2:27 AM, m zyzy wrote:
>
> Scott , my latest attempt to install in centos 5 this time work well by
> re-downloading .bin installer . But , in fc10 the
> ./postgresql-8.3.5-1-linux.bin command still to no avail.
> for now , the text mode ,
> ./postgresql-8.3.5-1-linux.bin --
Thank you Scott and Dave.
Dave , the url given , not really helpful unless I go through each and every
details,I dont know -searcg the page for the word segmentation returns
nothing . another thing ,just to let you know my postGIS installation
through StackBuilder still failed by returning error af
On Thu, Jan 15, 2009 at 7:47 AM, m zyzy wrote:
> I had this weird problem in CentOS 5 and Fedora 10 . the one-click binary
> installer failed
> execute this
> ./postgresql-8.3.5-1-linux.bin
>
> shows
>
> Segmentation fault
Do either of the suggestions at
http://bitrock.com/support_installbuilder_
On Thu, Jan 15, 2009 at 12:47 AM, m zyzy wrote:
> I had this weird problem in CentOS 5 and Fedora 10 . the one-click binary
> installer failed
> execute this
> ./postgresql-8.3.5-1-linux.bin
>
> shows
>
> Segmentation fault
I don't know what's causing it, I use the PGDG RHEL 5 packages for
pgsql
83 matches
Mail list logo
