Heikki Linnakangas wrote:
Jan Urbański wrote:
26763 3.5451 AllocSetCheck
Make sure you disable assertions before profiling.
Awww, darn. OK, here goes another set of results, without casserts this
time.
=== CVS HEAD ===
number of clients: 10
number of transactions per client: 10
On Wed, Aug 13, 2008 at 7:09 PM, Jaime Casanova wrote:
> any move in this?
I did some changes to pgbench in February and sent them to Andrew. No
reaction has been got so far.
--
Regards,
Michael Holzman
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to
On Thu, Jul 31, 2008 at 1:52 PM, claudio lezcano <[EMAIL PROTECTED]> wrote:
> Hi everybody
>
> Iam traying to build libpq.lib and libpq.dll library using Borland c++ 5.5
> and i got these error:
>
> "Error libpq.rc 1 11: Cannot open file: winver.h"
>
> I opened the libpq.rc file and i saw the refer
> "Decibel" == Decibel! <[EMAIL PROTECTED]> writes:
Decibel> OK, ran the test again via this query:
Decibel> explain analyze select (select value from oneblock where id = i)
Decibel> from generate_series(1,1) i, generate_series(1,10) j;
Decibel> changing 1,1 to 200,200 as needed. I
On Aug 13, 2008, at 20:12, Tom Lane wrote:
Wow. That sound awesome, Tom. Stupid question: Do these join types
have some sort of correspondence to the SQL standard?
Semi and anti joins are pretty standard concepts in relational theory,
but they have no direct mapping in the SQL join syntax. Yo
"David E. Wheeler" <[EMAIL PROTECTED]> writes:
> On Aug 13, 2008, at 17:31, Tom Lane wrote:
>> Introduce JOIN_SEMI and JOIN_ANTI join types,
> Wow. That sound awesome, Tom. Stupid question: Do these join types
> have some sort of correspondence to the SQL standard?
Semi and anti joins are prett
On Aug 13, 2008, at 17:31, Tom Lane wrote:
What's done:
Introduce JOIN_SEMI and JOIN_ANTI join types, the former replacing
JOIN_IN. Unify the InClauseInfo and OuterJoinInfo infrastructure into
"SpecialJoinInfo". Convert IN, EXISTS, and NOT EXISTS clauses at top
level of WHERE into semi and an
Uh, what version of Postgres are you testing? Would you email the
output of the compile so we can see why it isn't find it? My guess is
that borland CC isn't configured to find the Windows include directory.
---
claudio le
Decibel! <[EMAIL PROTECTED]> writes:
> It'd be useful to get strace data on this, but OS X doesn't have
> that :/ (and I'm on 10.4 so no dtrace either).
See ktrace.
regards, tom lane
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes
On Aug 13, 2008, at 3:52 PM, Decibel! wrote:
The problem is that by looking for a constant row, you're actually
eliminating the entire effect being tested, because the uncorrelated
subselect is run ONCE as an initplan, and the entire query time is
then spent elsewhere. The differences in runtime
Alvaro Herrera <[EMAIL PROTECTED]> writes:
> Tom Lane wrote:
>> It seems like we'll want to do it somehow. Perhaps the cleanest way is
>> to incorporate toast-table settings in the reloptions of the parent
>> table. Otherwise dump/reload is gonna be a mess.
> My question is whether there is inte
Tom Lane wrote:
> Alvaro Herrera <[EMAIL PROTECTED]> writes:
> > Does anyone see a need for having TOAST tables be individually
> > configurable for autovacuum? I've finally come around to looking at
> > being able to use ALTER TABLE for autovacuum settings, and I'm wondering
> > if we need to sup
Alvaro Herrera <[EMAIL PROTECTED]> writes:
> Does anyone see a need for having TOAST tables be individually
> configurable for autovacuum? I've finally come around to looking at
> being able to use ALTER TABLE for autovacuum settings, and I'm wondering
> if we need to support that case.
It seems
Gregory Stark <[EMAIL PROTECTED]> writes:
> That means going to the index meta page, find the fast root pointer, look up
> that page, look at the single leaf page pointer, look up that page, and do a
> binary search of the 200 leaf pointers. Once you find the resulting match,
> look up the heap pag
"Decibel!" <[EMAIL PROTECTED]> writes:
> Makes sense, and yeah, I was wondering a bit about that. I'll try to
> fake it out with offset 0 later on if no one beats me to it; I do still
> think we could just be seeing the effect of slogging through 200 tuples
> instead of going directly to the one w
Jan Urbański wrote:
26763 3.5451 AllocSetCheck
Make sure you disable assertions before profiling. Although I'm actually
a bit surprised the overhead isn't more than 3.5%, I've seen much higher
overheads on other tests, but it's still skewing the results.
- Heikki
--
Sent via pgsql-hack
Hackers and PG users,
Does anyone see a need for having TOAST tables be individually
configurable for autovacuum? I've finally come around to looking at
being able to use ALTER TABLE for autovacuum settings, and I'm wondering
if we need to support that case.
--
Alvaro Herrera
On Wed, Aug 13, 2008 at 07:33:40PM +0100, Andrew Gierth wrote:
> The following message is a courtesy copy of an article
> that has been posted to pgsql.hackers as well.
>
> > "Decibel!" == Decibel! <[EMAIL PROTECTED]> writes:
>
> Decibel> Roughly what I get on my MBP (I'm about a factor of
Heikki Linnakangas wrote:
Jan Urbański wrote:
through it. The only tiny ugliness is that there's one function used
for qsort() and another for bsearch(), because I'm sorting an array of
texts (from pg_statistic) and I'm binary searching for a lexeme
(non-NULL terminated string with length).
On Wed, 13 Aug 2008, Marko Kreen wrote:
On 8/8/08, Heikki Linnakangas <[EMAIL PROTECTED]> wrote:
One idea for fixing this is to make cancellation keys disposable, and
automatically issue a new one through the main connection when one is used,
but that's not completely trivial, and requires a ch
Jaime Casanova wrote:
On 1/21/08, Andrew Dunstan <[EMAIL PROTECTED]> wrote:
It's on my (very long) TODO list to add benchmarking as an option on the
buildfarm. If you're interested in working on it then contact me offline and
we will work on how to move forward.
any move in this?
Not much
On Wed, 2008-08-13 at 12:55 -0400, Bruce Momjian wrote:
> NTT is working with EnterpriseDB
> on the WAL steaming code to be released to the community.
Hopefully the code isn't steaming... :-)
and that we will be able to see detailed designs and code soon.
Might end up as a big pileup otherwise.
> "Decibel!" == Decibel! <[EMAIL PROTECTED]> writes:
Decibel> Roughly what I get on my MBP (I'm about a factor of 2
Decibel> slower). This makes me think it's an issue of having to slog
Decibel> through an entire page one row at a time vs just using the
Decibel> index. To test this I test
On 8/8/08, Heikki Linnakangas <[EMAIL PROTECTED]> wrote:
> It occurred to me a while ago that our query cancel messages are sent
> unencrypted, even when SSL is otherwise used. That's not a big issue on its
> own, because the cancellation message only contains the backend PID and the
> cancellation
Simon Riggs wrote:
>
> On Wed, 2008-08-13 at 11:27 -0400, Bruce Momjian wrote:
>
> > I disagree. If they make it the master they change the setting.
>
> It's not acceptable to force people to edit a configuration file when
> failover occurs. Some people wish to automate this and fumbling
> para
Decibel! wrote:
> On Aug 13, 2008, at 4:12 AM, Magnus Hagander wrote:
>> Tom Lane wrote:
>>> Decibel! <[EMAIL PROTECTED]> writes:
I disagree. While we don't guarantee stats are absolutely up-to-date,
or atomic I don't think that gives license for them to just magically
not exist some
Simon Riggs wrote:
>
> On Wed, 2008-08-13 at 11:17 -0400, Bruce Momjian wrote:
>
> > I think doing the WAL streaming and allowing a read-only slave is
> > enough work to keep Simon busy for quite some time. I don't
> > understand why the logical issue is being discussed at this stage ---
> > let
On Aug 12, 2008, at 4:52 PM, Andrew Gierth wrote:
"Tom" == Tom Lane <[EMAIL PROTECTED]> writes:
Proposal: Make the first block of a seq scan cost
random_page_cost, then after that every additional block costs
seq_page_cost.
?Tom> This is only going to matter for a table of 1 block (or at lea
On Wed, 2008-08-13 at 11:27 -0400, Bruce Momjian wrote:
> I disagree. If they make it the master they change the setting.
It's not acceptable to force people to edit a configuration file when
failover occurs. Some people wish to automate this and fumbling
parameter values at this important time
Ok, that worked!
Thank you very much, Jan and others who gave their input.
I did see Tom's input for the VARHDRSZ and tried that, but forgot to add
that again
when I called VARATT_SIZEP
Jan Urbański wrote:
Tim Hawes wrote:
@Jan:
It appears the cstring_to_text function is unique to the lates
Tim Hawes wrote:
@Jan:
It appears the cstring_to_text function is unique to the latest
PostgreSQL code. I do not have a def for that for PostgreSQL 8.2, and
Oh, I'm sorry, I forgot about that. cstring_to_text has been added only
recently (it's not even it 8.3, silly me).
Datum pl_master
On Wed, 2008-08-13 at 11:17 -0400, Bruce Momjian wrote:
> I think doing the WAL streaming and allowing a read-only slave is
> enough work to keep Simon busy for quite some time. I don't
> understand why the logical issue is being discussed at this stage ---
> let's get the other stuff done first
On 1/21/08, Andrew Dunstan <[EMAIL PROTECTED]> wrote:
>
> It's on my (very long) TODO list to add benchmarking as an option on the
> buildfarm. If you're interested in working on it then contact me offline and
> we will work on how to move forward.
>
any move in this?
--
regards,
Jaime Casanova
On Aug 13, 2008, at 4:12 AM, Magnus Hagander wrote:
Tom Lane wrote:
Decibel! <[EMAIL PROTECTED]> writes:
I disagree. While we don't guarantee stats are absolutely up-to-
date,
or atomic I don't think that gives license for them to just
magically
not exist sometimes.
Would it really be tha
> > > Proposal: Make the first block of a seq scan cost random_page_cost, then
> > > after that every additional block costs seq_page_cost.
+1
> AFAICS the cost cross-over is much higher than the actual elapsed time
> cross-over for both narrow and wide tables.
Which makes absolute sense, since
Simon Riggs wrote:
> > > The main point of the post is that the parameter would be transaction
> > > controlled, so *must* be set in the transaction and thus *must* be set
> > > on the master. Otherwise the capability is not available in the way I am
> > > describing.
> >
> > Oh, so synchronous_co
Simon Riggs wrote:
>
> On Wed, 2008-08-13 at 15:38 +0200, Markus Wanner wrote:
>
> > Simon Riggs wrote:
> > > Classification of Replication Techniques
> >
> > Thanks for your classifications. It helps a great deal to clarify.
> >
> > > Type 2 is where you ship the WAL (efficient) then use it to
>>> Tom Lane <[EMAIL PROTECTED]> wrote:
> BTW, should we make all of this conditional on the use of an SSL
> connection? If the original sending of the cancel key isn't secure
> against sniffing, it's hard to see what anyone is buying with all
the
> added computation.
+1
All of our importan
Added to TODO:
> o Add CREATE TABLE LIKE ... INCLUDING COMMENTS
> * Allow table and index WITH options to be specified via hooks, for use
> with plugins like GiST index methods
---
ITAGAKI Takahiro wrote:
>
> Sim
Thank you for your replies, however, it still is not working, see below...
Andrew Chernow wrote:
Tim Hawes wrote:
text * pl_masterkey(PG_FUNCTION_ARGS)
{
char *e_var = getenv("PGMASTERKEY");
size_t length = VARSIZE(e_var) - VARHDRSZ;
The VARSIZE macro is for variable length structures, l
Hi there... Configuration autotuning is something I am really interested in.
I have seen this page: http://wiki.postgresql.org/wiki/GUCS_Overhaul and
a couple of emails mentioning this, so I wanted to ask is someone already
on it? If yes, I'd like to contribute.
Ideally, an external little app sho
Andrew Chernow <[EMAIL PROTECTED]> writes:
> Tim Hawes wrote:
>> char *e_var = getenv("PGMASTERKEY");
>> size_t length = VARSIZE(e_var) - VARHDRSZ;
> The VARSIZE macro is for variable length structures, like a text or
> bytea which contains a length and data member. You are using this macro
> o
Tom Lane wrote:
> Magnus Hagander <[EMAIL PROTECTED]> writes:
>> Tom Lane wrote:
>>> * Server accepts two different styles of cancel messages, identified
>>> by different protocol numbers.
>
>> With the additional point that there is a GUC variable to turn this off
>> or warn about it, right?
>
>
Magnus Hagander <[EMAIL PROTECTED]> writes:
> Tom Lane wrote:
>> * Server accepts two different styles of cancel messages, identified
>> by different protocol numbers.
> With the additional point that there is a GUC variable to turn this off
> or warn about it, right?
I see pretty much no value i
Tom Lane wrote:
> Magnus Hagander <[EMAIL PROTECTED]> writes:
>> What would work is using a parameter field, per Stephen's suggestion
>> elsewhere in the thread. Older libpq versions should just ignore the
>> parameter if they don't know what it is.
>
> +1 for that one; we have done it already for
Magnus Hagander <[EMAIL PROTECTED]> writes:
> What would work is using a parameter field, per Stephen's suggestion
> elsewhere in the thread. Older libpq versions should just ignore the
> parameter if they don't know what it is.
+1 for that one; we have done it already for several send-at-startup
Stephen R. van den Berg wrote:
> Magnus Hagander wrote:
>> Gregory Stark wrote:
>>> "Magnus Hagander" <[EMAIL PROTECTED]> writes:
>>> We could have the server indicate it's the new protocol by sending the
>>> initial
>>> cancel key twice. If the client sees more than one cancel key it
>>> automat
On Wed, 2008-08-13 at 15:38 +0200, Markus Wanner wrote:
> Simon Riggs wrote:
> > Classification of Replication Techniques
>
> Thanks for your classifications. It helps a great deal to clarify.
>
> > Type 2 is where you ship the WAL (efficient) then use it to reconstruct
> > SQL (flexible) and t
Magnus Hagander wrote:
>Gregory Stark wrote:
>> "Magnus Hagander" <[EMAIL PROTECTED]> writes:
>> We could have the server indicate it's the new protocol by sending the
>> initial
>> cancel key twice. If the client sees more than one cancel key it
>> automatically
>> switches to new-style cancel m
Hi,
Simon Riggs wrote:
Classification of Replication Techniques
Thanks for your classifications. It helps a great deal to clarify.
Type 2 is where you ship the WAL (efficient) then use it to reconstruct
SQL (flexible) and then apply that to other nodes. It is somewhat harder
than type 1, but
Magnus Hagander wrote:
> Tom Lane wrote:
>> Decibel! <[EMAIL PROTECTED]> writes:
>>> I disagree. While we don't guarantee stats are absolutely up-to-date,
>>> or atomic I don't think that gives license for them to just magically
>>> not exist sometimes.
>>> Would it really be that hard to have
On Wed, 2008-08-13 at 11:27 +0200, Markus Wanner wrote:
> Hi,
>
> Robert Hodges wrote:
> > Part of this is semantics—I like Simon’s logical vs. physical
> > terminology because it distinguishes neatly between replication that
> > copies implementation down to OIDs etc. and replication that copi
Hello.
Here are these functions with detailed documentation:
http://en.dklab.ru/lib/dklab_postgresql_patch/
intagg.int_array_append_aggregate(int[]): fast merge arrays into one large
list
intarray._int_group_count_sort(int[], bool): frequency-based sorting
intarray.bidx(int[], int): binary search
Tim Hawes wrote:
text * pl_masterkey(PG_FUNCTION_ARGS)
{
char *e_var = getenv("PGMASTERKEY");
size_t length = VARSIZE(e_var) - VARHDRSZ;
The VARSIZE macro is for variable length structures, like a text or
bytea which contains a length and data member. You are using this macro
on a regul
Gregory Stark wrote:
> "Magnus Hagander" <[EMAIL PROTECTED]> writes:
>
>> Yeah, that's the point that will require a protocol bump, I think. Since
>> there is no response to the cancel packet, we can't even do things like
>> sending in a magic key and look at the response (which would be a rather
Tim Hawes wrote:
Hello all,
I am trying to write an extension in C that returns a simple environment
variable. The code compiles without any complaint or warning, and it
loads fine into the database, however, when I run the function, I get
disconnected from the server.
Here is my C code:
#
Hello all,
I am trying to write an extension in C that returns a simple environment
variable. The code compiles without any complaint or warning, and it
loads fine into the database, however, when I run the function, I get
disconnected from the server.
Here is my C code:
#include
#include
"Magnus Hagander" <[EMAIL PROTECTED]> writes:
> Yeah, that's the point that will require a protocol bump, I think. Since
> there is no response to the cancel packet, we can't even do things like
> sending in a magic key and look at the response (which would be a rather
> ugly hack, but doable if w
Tom Lane wrote:
>[ thinks for a bit... ] You could make it a change in the cancel
>protocol, which is to some extent independent of the main FE/BE
>protocol. The problem is: how can the client know whether it's okay to
>use this new protocol for cancel?
Two options:
a. Send two cancelkeys in rap
On Tue, 2008-08-12 at 23:22 -0400, Tom Lane wrote:
> Gregory Stark <[EMAIL PROTECTED]> writes:
> >> On Tue, 2008-08-12 at 15:46 -0400, Tom Lane wrote:
> >>> This is only going to matter for a table of 1 block (or at least very
> >>> few blocks), and for such a table it's highly likely that it's in
Hi,
Robert Hodges wrote:
Part of this is semantics—I like Simon’s logical vs. physical
terminology because it distinguishes neatly between replication that
copies implementation down to OIDs etc. and replication that copies data
content including schema changes but not implementation.
So far
Tom Lane wrote:
> Magnus Hagander <[EMAIL PROTECTED]> writes:
>> Andrew Gierth wrote:
>>> That's easily solved: when the client wants to do a cancel, have it
>>> send, in place of the actual cancel key, an integer N and the value
>>> HMAC(k,N) where k is the cancel key. Replay is prevented by requi
Tom Lane wrote:
> Decibel! <[EMAIL PROTECTED]> writes:
>> I disagree. While we don't guarantee stats are absolutely up-to-date,
>> or atomic I don't think that gives license for them to just magically
>> not exist sometimes.
>
>> Would it really be that hard to have the system copy the file ou
On Tue, 2008-08-12 at 13:33 -0400, Bruce Momjian wrote:
> Simon Riggs wrote:
> > > > > with values of:
> > > > >
> > > > > nothing:have network traffic send WAL as needed
> > > > > netflush: wait for flush of WAL network packets to slave
> > > > > process:w
Hi,
Alvaro Herrera wrote:
Actually I think the idea here is to take certain WAL records, translate
them into "portable" constructs, ship them,
At which point it clearly shouldn't be called a WAL shipping method.
What would it have to do with the WAL at all, then? Why translate from
WAL reco
Hi,
Robert Hodges wrote:
I like Simon’s logical vs. physical terminology
So far, it seems to mainly have caused confusion (physical replication,
but logical application? logical replication using WAL shipping?). At
least I still prefer the more meaningful and descriptive terms, like
"log sh
66 matches
Mail list logo
