Martijn van Oosterhout írta:
> On Sat, Apr 10, 2010 at 02:36:41PM +0200, Boszormenyi Zoltan wrote:
>
The above is quite reproducable, "pg_ctl stop -m immediate"
"usually" inflated my serial sequence, but I had two occasions
when not. The 69 -> 70 was one. The inflated increase is
On Sat, Apr 10, 2010 at 8:23 AM, Erik Rijkers wrote:
> I understand that in the scale=1000 case, there is a huge
> cache effect, but why doesn't that apply to the pgbench runs
> against the standby? (and for the scale=10_000 case the
> differences are still rather large)
I guess that this perfor
Fujii Masao wrote:
> doc/src/sgml/config.sgml
> -archival or to recover from a checkpoint. If standby_keep_segments
> +archival or to recover from a checkpoint. If
> standby_keep_segments
>
> The word "standby_keep_segments" always needs the tag, I think.
Thanks, fixed.
> We sho
Jaime Casanova wrote:
> On Fri, Apr 9, 2010 at 3:39 PM, Jaime Casanova
> wrote:
>> but, my main concern is why it was asking for
>> "00010006"? is this normal? is this standby's way of
>> saying i'm working but i have nothing to do?
Yes.
>> when that happens after a standby resta
On Mon, Apr 12, 2010 at 2:16 AM, Pavel Stehule wrote:
> I am sure so
> dynamical materialised views is bad task for GSoC - it is too large,
> too complex. Manually refreshed views is adequate to two months work
> and it has sense.
That is my feeling also - though I fear that even the simplest
pos
On Mon, Apr 12, 2010 at 5:06 AM, Fujii Masao wrote:
> On Sat, Apr 10, 2010 at 8:23 AM, Erik Rijkers wrote:
>> I understand that in the scale=1000 case, there is a huge
>> cache effect, but why doesn't that apply to the pgbench runs
>> against the standby? (and for the scale=10_000 case the
>> di
On Thu, Apr 8, 2010 at 5:01 PM, Fujii Masao wrote:
>> If it does, there should be
>> some way to get PGXS to execute that rule as well, I'm sure.
>
> If we can copy/link the source file defining "new PQexec" when
> we compile the dblink, DLL doesn't seem to be required. So I
> stop creating new DL
On Sat, April 10, 2010 01:23, Erik Rijkers wrote:
> Using 9.0devel cvs HEAD, 2010.04.08.
>
> I am trying to understand the performance difference
> between primary and standby under a standard pgbench
> read-only test.
>
> server has 32 GB, 2 quadcores.
>
> primary:
> tps = 34606.747930 (includin
On Mon, Apr 12, 2010 at 7:07 AM, Robert Haas wrote:
> On Mon, Apr 12, 2010 at 5:06 AM, Fujii Masao wrote:
>> On Sat, Apr 10, 2010 at 8:23 AM, Erik Rijkers wrote:
>>> I understand that in the scale=1000 case, there is a huge
>>> cache effect, but why doesn't that apply to the pgbench runs
>>> aga
On Mon, Apr 12, 2010 at 7:41 PM, Heikki Linnakangas
wrote:
>> We should remove the document "25.2.5.2. Monitoring"?
>
> I updated it to no longer claim that the primary can run out of disk
> space because of a hung WAL sender. The information about calculating
> the lag between primary and standby
On Mon, April 12, 2010 14:22, Erik Rijkers wrote:
> On Sat, April 10, 2010 01:23, Erik Rijkers wrote:
Oops, typos in that pseudo loop:
of course there was a pgbench init step after that first line.
> for scale in 10 100 500 1000
pgbench ... # initialise
sleep ((scale / 10) * 60)
>
On Mon, Apr 12, 2010 at 6:41 AM, Heikki Linnakangas
wrote:
>> Why is standby_keep_segments used even if max_wal_senders is zero?
>> In that case, ISTM we don't need to keep any WAL files in pg_xlog
>> for the standby.
>
> True. I don't think we should second guess the admin on that, though.
> Perh
On Mon, Apr 12, 2010 at 8:32 AM, Jim Mlodgenski wrote:
>> I think we need to investigate this more. It's not going to look good
>> for the project if people find that a hot standby server runs two
>> orders of magnitude slower than the primary.
> As a data point, I did a read only pgbench test an
* Robert Haas [100412 07:10]:
> I think we need to investigate this more. It's not going to look good
> for the project if people find that a hot standby server runs two
> orders of magnitude slower than the primary.
Yes, it's not "good", but it's a known problem. We've had people
complaining
And I see now that he's doing a stream of read-only queries on a slave,
presumably with no WAL even being replayed...
Sorry for the noise
a.
* Aidan Van Dyk [100412 09:40]:
> * Robert Haas [100412 07:10]:
>
> > I think we need to investigate this more. It's not going to look good
> > fo
>Aidan Van Dyk wrote:
> We've had people complaining that wal-replay can't keep up with a
> wal stream from a heavy server.
I thought this thread was about the slow performance running a mix
of read-only queries on the slave versus the master, which doesn't
seem to have anything to do with the
resending this message, as it seems to have bounced.
(below, I did fix the typo in the pseudocode loop)
Original Message
Subject: Re: [HACKERS] testing HS/SR - 1 vs 2 performance
From:"Erik Rijkers"
Date:M
On Mon, Apr 12, 2010 at 1:48 AM, Fujii Masao wrote:
> On Mon, Apr 12, 2010 at 3:29 PM, Jaime Casanova
> wrote:
>> On Mon, Apr 12, 2010 at 1:21 AM, Fujii Masao wrote:
>>> Didn't the standby
>>> accept connections before executing pgbench?
>>>
>>
>> nop, and last time i try it was in that state fo
The patch I sent earlier is flaud with respect to subplan parameter
numbering, I counted from zero where the parParam list had to be used.
Yeb Havinga wrote:
See patch below against HEAD.
Example of query against catalog:
postgres=# explain verbose select oid::int + 1,(select oid from
pg_cl
Fujii Masao wrote:
> If adding new shared library is too big change at this point, I think
> that we should postpone the fix only for dblink to 9.1 or later. Since
> no one has complained about this long-term problem of dblink, I'm not
> sure it really should be fixed right now. Thought?
I would a
I just did a checkout from HEAD (a few minutes ago) and ran this:
make distclean ; ./configure --prefix=/usr/local/pgsql-serializable
--enable-integer-datetimes --enable-debug --enable-cassert
--enable-depend --with-libxml && make check
I got a failure on the random test. Unfortunately I didn'
On Mon, Apr 12, 2010 at 9:27 AM, Jaime Casanova
wrote:
> On Mon, Apr 12, 2010 at 1:48 AM, Fujii Masao wrote:
>> On Mon, Apr 12, 2010 at 3:29 PM, Jaime Casanova
>> wrote:
>>> On Mon, Apr 12, 2010 at 1:21 AM, Fujii Masao wrote:
Didn't the standby
accept connections before executing pgbe
On Mon, Apr 12, 2010 at 13:54, Fujii Masao wrote:
> On Thu, Apr 8, 2010 at 5:01 PM, Fujii Masao wrote:
>>> If it does, there should be
>>> some way to get PGXS to execute that rule as well, I'm sure.
>>
>> If we can copy/link the source file defining "new PQexec" when
>> we compile the dblink, DL
On 4/9/10 1:36 PM, pavelbaros wrote:
> 2) change rewriter
> - usually, view is relation with defined rule and when rewriting, rule
> is fired and relation (view) is replaced by definition of view. If
> relation do not have rule, planner and executor behave to it as physical
> table (relation). In c
On Mon, Apr 12, 2010 at 1:50 PM, Josh Berkus wrote:
> On 4/9/10 1:36 PM, pavelbaros wrote:
>> 2) change rewriter
>> - usually, view is relation with defined rule and when rewriting, rule
>> is fired and relation (view) is replaced by definition of view. If
>> relation do not have rule, planner and
Josh Berkus wrote:
There are basically 2 major parts for materialized views:
A) Planner: Getting the query planner to swap in the MatView for part of
a query automatically for query plan portions which the MatView supports;
B) Maintenance: maintaining the MatView data according to the programmed
On Mon, Apr 12, 2010 at 3:43 PM, Greg Smith wrote:
> Josh Berkus wrote:
>>
>> There are basically 2 major parts for materialized views:
>> A) Planner: Getting the query planner to swap in the MatView for part of
>> a query automatically for query plan portions which the MatView supports;
>> B) Mai
On 4/10/10 7:00 AM, Jean-Gérard Pailloncy wrote:
> Hello,
>
> 1) VPD: Virtual Private Database
> I would appreciate to have a new feature in PostgreSQL.
> This is an oracle-like feature that implement "Row Level Security".
> This feature may be emulated by using VIEW/RULE but this is very time
> c
Greg,
> I'm not saying someone can't jump right into (3), using the current
> implementations for (1) and (2) that are floating around out there. I
> just think it would end up wasting a fair amount of work on prototypes
> that don't work quite the same way as the eventual fully integrated
> vers
Josh Berkus wrote:
What would be the use case for (1) by itself?
There isn't any use case for just working on the infrastructure, just
like there's no use case for "Syntax for partitioning" on its own. That
why people rarely work on that part of these problems--it's boring and
produces n
FYI, I think Michael Meskes applied this patch, though I didn't see you
emailed that it was applied.
---
Boszormenyi Zoltan wrote:
> Hi,
>
> here's a little beautified patch:
> - more logical parameter order in ECPGdump_a_t
On Tue, Apr 13, 2010 at 1:56 AM, Magnus Hagander wrote:
>> If adding new shared library is too big change at this point, I think
>> that we should postpone the fix only for dblink to 9.1 or later. Since
>> no one has complained about this long-term problem of dblink, I'm not
>> sure it really shou
Simon Riggs wrote:
> On Sun, 2010-04-04 at 22:12 -0400, Bruce Momjian wrote:
> > Simon Riggs wrote:
> > >
> > > How about we call it "exclusivity constraints".
> > >
> > > Not much of a change, but helps to differentiate.
> >
> > Well, the keyword is EXCLUDE so we could call it "EXCLUDE contrain
> I don't want to see Materialized Views wander down the same path as
> partitioning, where lots of people produce "fun parts" patches, while
> ignoring the grunt work of things like production quality catalog
> support for the feature. I think Pavel's proposal got that part right
> by starting w
On Mon, Apr 12, 2010 at 11:27 PM, Jaime Casanova
wrote:
>> 1. start the primary
>> 2. pg_start_backup()
>> 3. copy $PGDATA from the primary to the standby
>> 4. pg_stop_backup();
>> 5. create the recovery.conf and start the standby
>
> execute some WAL-logged action (i've seen this happen even wit
On Mon, Apr 12, 2010 at 9:28 PM, Bruce Momjian wrote:
> Simon Riggs wrote:
>> On Sun, 2010-04-04 at 22:12 -0400, Bruce Momjian wrote:
>> > Simon Riggs wrote:
>> > >
>> > > How about we call it "exclusivity constraints".
>> > >
>> > > Not much of a change, but helps to differentiate.
>> >
>> > Well
Robert Haas wrote:
> On Mon, Apr 12, 2010 at 9:28 PM, Bruce Momjian wrote:
> > Simon Riggs wrote:
> >> On Sun, 2010-04-04 at 22:12 -0400, Bruce Momjian wrote:
> >> > Simon Riggs wrote:
> >> > >
> >> > > How about we call it "exclusivity constraints".
> >> > >
> >> > > Not much of a change, but hel
On Mon, Apr 12, 2010 at 11:03 PM, Bruce Momjian wrote:
> Robert Haas wrote:
>> On Mon, Apr 12, 2010 at 9:28 PM, Bruce Momjian wrote:
>> > Simon Riggs wrote:
>> >> On Sun, 2010-04-04 at 22:12 -0400, Bruce Momjian wrote:
>> >> > Simon Riggs wrote:
>> >> > >
>> >> > > How about we call it "exclusivi
Hello:
I am brand new to Postgresql.
I ran the following commands.
./configure
gmake
su
gmake install
adduser postgres
mkdir /usr/local/pgsql/data
chown postgres /usr/local/pgsql/data
su - postgres
/usr/local/pgsql/bin/initdb -D /usr/local/pgsql/data
/usr/local/pgsql/bin/postmaster -D /usr/local/
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Mon, Apr 12, 2010 at 08:31:38PM -0700, Murali M. Krishna wrote:
> Hello:
>
> I am brand new to Postgresql.
>
> I ran the following commands.
> ./configure
> gmake
> su
> gmake install
> adduser postgres
> mkdir /usr/local/pgsql/data
> chown postgr
Yes, he applied the first version without seeing this one,
then he asked for a re-diff privately.
Bruce Momjian írta:
> FYI, I think Michael Meskes applied this patch, though I didn't see you
> emailed that it was applied.
>
> ---
Josh Berkus wrote:
I just worry about any feature which doesn't get as far as a
user-visible implementation. If someone doesn't do the rest of the
parts soon, such features tend to atrophy because nobody is using them.
While they're limited, there are complexly viable prototype quality
imp
The OS is
Fedora 12.
-
Please visit NumberFest.com for educational number puzzles & mind exercises for
all ages! And please tell your friends about it. Thank You!
--- On Mon, 4/12/10, to...@tuxteam.de wrote:
From: to...@tux
I could reproduce this on my laptop, standby is about 20% slower. I ran
oprofile, and what stands out as the difference between the master and
standby is that on standby about 20% of the CPU time is spent in
hash_seq_search(). The callpath is GetSnapshotDat() ->
KnownAssignedXidsGetAndSetXmin() ->
"Kevin Grittner" writes:
> I got a failure on the random test.
This used to be common. Peter tweaked the test a few years ago to
reduce the probability of failure, but IIRC it's still not zero (and
probably can't be made zero without rendering the test meaningless).
I think most likely you just
On Mon, Apr 12, 2010 at 8:32 PM, Fujii Masao wrote:
> On Mon, Apr 12, 2010 at 11:27 PM, Jaime Casanova
> wrote:
>>> 1. start the primary
>>> 2. pg_start_backup()
>>> 3. copy $PGDATA from the primary to the standby
>>> 4. pg_stop_backup();
>>> 5. create the recovery.conf and start the standby
>>
>
Robert Haas writes:
> On Mon, Apr 12, 2010 at 11:03 PM, Bruce Momjian wrote:
>> Fine, then we will just have to live with "exclusion constraints" and
>> "contraint exclusion".
> I am not necessarily 100% averse to changing it... just saying that it
> shouldn't be done unless we have a clear cons
47 matches
Mail list logo