I won't have access to the original testcase and server for a few days
but I just redid some testing on a slower personal box of mine with a
smaller(but similiar) testset and on that box I could not reproduce that
issue.
So the problem is either caused by the size of the table or somehow by
the
On Wed, Jun 28, 2006 at 01:09:23PM +0400, Teodor Sigaev wrote:
We tested gin with 5 millions records (but not a wikipedia's text, but with
blog records which is usually mush shorter) and index creation time was
near 8 hours on Dual Xeon 3.2, but we didn't try GiST. In other hand,
creating
On Wed, 28 Jun 2006, Martijn van Oosterhout wrote:
On Wed, Jun 28, 2006 at 01:09:23PM +0400, Teodor Sigaev wrote:
We tested gin with 5 millions records (but not a wikipedia's text, but with
blog records which is usually mush shorter) and index creation time was
near 8 hours on Dual Xeon 3.2,
Tom did commit a patch a while ago which made a huge difference in
index creation time for tsearch by changing one routine. I don't know
if it got backpatched, so it might be worth checking people are working
on the same version.
I saw that patch, but I still think that 7 minutes is too small
Hi,
I need help finding out how to determine if two types are equality compatible
and compare them.
I am using the following call to check for equality between two values:
DatumGetBool(\
FunctionCall2((fctx-tupleSetAttEQFunctions[attID]-eq_opr_finfo)\
, lvalue, rvalue))
The
Hi,
From: Tom Lane [EMAIL PROTECTED]
Subject: Re: [HACKERS] SO_SNDBUF size is small on win32?
Date: Tue, 27 Jun 2006 14:43:57 -0400
What I would think might help is a patch on the libpq side (because it
*does* use a nonblocking socket) to avoid sending more than
8K per WSASend call.
Yoshiyuki Asaba [EMAIL PROTECTED] writes:
I think libpq interface does not use non-blocking socket.
Not unless the Windows port has also disabled pg_set_noblock ...
regards, tom lane
---(end of broadcast)---
TIP 3: Have
On Wed, Jun 28, 2006 at 03:25:57PM +0300, Tzahi Fadida wrote:
Hi,
I need help finding out how to determine if two types are equality compatible
and compare them.
snip
Currently i only allow two values only of the same type but i wish to allow
to compare values like 20.2=?20.2 or 20=?20.0
I have an optimization I'd like to see which I think should be pretty
easy for someone familiar with the planner code to implement. My
situation is this: I have an application using veil[1]. Essentially, I
have a schema private and another public. Private contains regular
tables, where private
On Wed, Jun 28, 2006 at 10:35:37AM -0400, Phil Frost wrote:
I have an optimization I'd like to see which I think should be pretty
easy for someone familiar with the planner code to implement. My
situation is this: I have an application using veil[1]. Essentially, I
have a schema private and
Martijn van Oosterhout kleptog@svana.org writes:
On Wed, Jun 28, 2006 at 03:25:57PM +0300, Tzahi Fadida wrote:
I need help finding out how to determine if two types are equality compatible
and compare them.
Fortunatly the backend contains functions that do all this already.
Check out
On Wed, Jun 28, 2006 at 05:11:59PM +0200, Martijn van Oosterhout wrote:
On Wed, Jun 28, 2006 at 10:35:37AM -0400, Phil Frost wrote:
I have an optimization I'd like to see which I think should be pretty
easy for someone familiar with the planner code to implement. My
situation is this: I
Phil Frost [EMAIL PROTECTED] writes:
The planner in fact can move the function around without changing the
output.
Not when it's within the nullable side of an outer join --- moving a
WHERE clause up out of that would make the difference between no row
out, and a null-extended row out, which
From: Tom Lane [EMAIL PROTECTED]
Subject: Re: [HACKERS] SO_SNDBUF size is small on win32?
Date: Wed, 28 Jun 2006 09:54:21 -0400
Yoshiyuki Asaba [EMAIL PROTECTED] writes:
I think libpq interface does not use non-blocking socket.
Not unless the Windows port has also disabled pg_set_noblock
Buildfarm member platypus is showing a regression failure that I'm
surprised we have not seen before:
http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=platypusdt=2006-06-28%2014:05:01
Basically what this is showing is that when there is more than one
referencing table, the order in which things
On Wed, Jun 28, 2006 at 11:40:52AM -0400, Tom Lane wrote:
Phil Frost [EMAIL PROTECTED] writes:
The planner in fact can move the function around without changing the
output.
Not when it's within the nullable side of an outer join --- moving a
WHERE clause up out of that would make the
We have now experienced index corruption on two separate but identical
slony clusters. In each case the slony subscriber failed after
attempting to insert a duplicate record. In each case reindexing the
sl_log_1 table on the provider fixed the problem.
The latest occurrence was on our
Martijn van Oosterhout wrote:
On Wed, Jun 28, 2006 at 01:56:47PM +0200, Thomas Hallgren wrote:
A user that is trusted with installing a C-function in the backend is
free to scan the process memory anyway so in what way did that increase
the security? IMHO, the only relevant security in that
Tom Lane wrote:
1. Find a way to make the processing order consistent (eg by driving it
off OID ordering). Doesn't seem easy, but maybe I'm missing an idea.
Hmm, what about
1. get the complete list of tables to truncate, AccessShareLock'ed, get
their names
2. release locks
3. sort the list
On Wed, Jun 28, 2006 at 06:39:16PM +0200, Thomas Hallgren wrote:
Instead, I would like to humbly request the inclusion of a UUID datatype
(or an opaque 128 bit datatype) in the core package. It's increasingly
ISTM that we get enough requests for this that it's probably worth
doing.
--
Jim C.
Alvaro Herrera [EMAIL PROTECTED] writes:
Tom Lane wrote:
1. Find a way to make the processing order consistent (eg by driving it
off OID ordering). Doesn't seem easy, but maybe I'm missing an idea.
Hmm, what about
1. get the complete list of tables to truncate, AccessShareLock'ed, get
On Wed, Jun 28, 2006 at 06:39:16PM +0200, Thomas Hallgren wrote:
Instead, I would like to humbly request the inclusion of a UUID datatype
(or an opaque 128 bit datatype) in the core package. It's increasingly
common and some databases (MS SQLServer) already have built in support for
it.
We
On Wed, Jun 28, 2006 at 01:13:42PM -0400, Tom Lane wrote:
One thing I was toying with was to add an index to pg_constraint on,
say, (confrelid, conrelid), and to replace the existing seqscans for FK
constraints with scans using this index. The second-column ordering
would guarantee everybody
On Wed, June 28, 2006 1:14 pm, [EMAIL PROTECTED] wrote:
On Wed, Jun 28, 2006 at 06:39:16PM +0200, Thomas Hallgren wrote:
Instead, I would like to humbly request the inclusion of a UUID
datatype (or an opaque 128 bit datatype) in the core package. It's
increasingly common and some databases
Marc Munro [EMAIL PROTECTED] writes:
We have now experienced index corruption on two separate but identical
slony clusters. In each case the slony subscriber failed after
attempting to insert a duplicate record. In each case reindexing the
sl_log_1 table on the provider fixed the problem.
On Wed, Jun 28, 2006 at 01:18:39PM -0400, A.M. wrote:
On Wed, June 28, 2006 1:14 pm, [EMAIL PROTECTED] wrote:
On Wed, Jun 28, 2006 at 06:39:16PM +0200, Thomas Hallgren wrote:
Instead, I would like to humbly request the inclusion of a UUID
datatype (or an opaque 128 bit datatype) in the
Tom Lane wrote:
Buildfarm member platypus is showing a regression failure that I'm
surprised we have not seen before:
http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=platypusdt=2006-06-28%2014:05:01
Basically what this is showing is that when there is more than one
referencing table, the
On Wed, Jun 28, 2006 at 12:38:50PM -0500, Jim C. Nasby wrote:
On Wed, Jun 28, 2006 at 01:18:39PM -0400, A.M. wrote:
On Wed, June 28, 2006 1:14 pm, [EMAIL PROTECTED] wrote:
On Wed, Jun 28, 2006 at 06:39:16PM +0200, Thomas Hallgren wrote:
Instead, I would like to humbly request the
Jim C. Nasby wrote:
On Wed, Jun 28, 2006 at 01:18:39PM -0400, A.M. wrote:
On Wed, June 28, 2006 1:14 pm, [EMAIL PROTECTED] wrote:
On Wed, Jun 28, 2006 at 06:39:16PM +0200, Thomas Hallgren wrote:
Instead, I would like to humbly request the inclusion of a UUID
datatype (or an
Tom Lane wrote:
Alvaro Herrera [EMAIL PROTECTED] writes:
Tom Lane wrote:
1. Find a way to make the processing order consistent (eg by driving it
off OID ordering). Doesn't seem easy, but maybe I'm missing an idea.
Hmm, what about
1. get the complete list of tables to truncate,
[EMAIL PROTECTED] wrote:
I've downloaded the version off pgfoundry.org. It is broken. It leaks
memory, and if memory is correct it can cause the client to core dump.
Also it couldn't possibly be included in core, since it's based on a
GPL'ed UUID library. If you have a more appropiately
Andrew Dunstan [EMAIL PROTECTED] writes:
If this were a significant risk wouldn't we have seen many such failures
before now?
Hard to tell. It's possibly architecture-dependent, for one thing
(MAXALIGN will affect space availability). Since this happened in a
parallel regression run, it
Alvaro Herrera wrote:
Tom Lane wrote:
Alvaro Herrera [EMAIL PROTECTED] writes:
Tom Lane wrote:
1. Find a way to make the processing order consistent (eg by driving it
off OID ordering). Doesn't seem easy, but maybe I'm missing an idea.
Hmm, what about
1. get the complete
Alvaro Herrera [EMAIL PROTECTED] writes:
Tom Lane wrote:
One thing I was toying with was to add an index to pg_constraint on,
say, (confrelid, conrelid), and to replace the existing seqscans for FK
constraints with scans using this index.
I think there is more than one place that would
bruce wrote:
Greg Stark wrote:
Bruce Momjian [EMAIL PROTECTED] writes:
PFC wrote:
My idea is that if an UPDATE places the new tuple on the same page as
the old tuple, it will not create new index entries for any indexes
where the key doesn't change.
On Wed, Jun 28, 2006 at 01:49:55PM -0400, Andrew Dunstan wrote:
Personally I don't buy the misuse objection - we already have plenty of
things that can be misused. As long as there is a reasonable valid use
and we can make it portable enough, I think there is a good case for
including it.
Bruce Momjian [EMAIL PROTECTED] writes:
Here is an overview of the SITC method:
http://momjian.us/cgi-bin/pgsitc
A pretty fundamental problem is that the method assumes it's OK to
change the CTID of a live tuple (by swapping its item pointer with some
expired version). It is not --- this
Andrew Dunstan [EMAIL PROTECTED] writes:
Tom Lane wrote:
Buildfarm member platypus is showing a regression failure that I'm
surprised we have not seen before:
http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=platypusdt=2006-06-28%2014:05:01
If this were a significant risk wouldn't we have
I wrote:
Alvaro Herrera [EMAIL PROTECTED] writes:
Tom Lane wrote:
One thing I was toying with was to add an index to pg_constraint on,
say, (confrelid, conrelid), and to replace the existing seqscans for FK
constraints with scans using this index.
I think there is more than one place that
Tom Lane [EMAIL PROTECTED] writes:
Martijn van Oosterhout kleptog@svana.org writes:
The input functions get it, the output functions (bpcharout,
bpcharsend, etc) don't. Which makes it kind of hard to print a raw
value if you don't know how long it's going to be. They used to, but
that
Tom Lane [EMAIL PROTECTED] writes:
Phil Frost [EMAIL PROTECTED] writes:
The planner in fact can move the function around without changing the
output.
Not when it's within the nullable side of an outer join --- moving a
WHERE clause up out of that would make the difference between no row
Greg Stark [EMAIL PROTECTED] writes:
Hm, so it could be stored on disk without the length header as long as
the length header is added to the in-memory representation? I don't
think the type system has hooks for reading and storing data to disk
though.
No, it doesn't, and we'd pay a nonzero
42 matches
Mail list logo