On Tue, 2007-03-06 at 00:18 -0500, Tom Lane wrote:
Sounds like #ifdef time to me --- but it seems a bit strange; wouldn't
the Python guys have taken a bit more care for compatibility of
user-supplied code?
Yeah, I was a bit surprised as well. I won't claim to have any
familiarity with the
On Tue, 2007-03-06 at 00:54 +0100, Florian G. Pflug wrote:
Simon Riggs wrote:
But it would break the idea of letting a second seqscan follow in the
first's hot cache trail, no?
No, but it would make it somewhat harder to achieve without direct
synchronization between scans. It could still
I need some help in building new index structure using GIST.
I have to include a procedure for restructuring the tree immediately after
i have created the index using gist. How do i do it?
Actually I am trying to implement the slim-tree index using gist. To
increase the performance, a slim-down
Mike Rylander wrote:
The patch adds support for default XML namespaces in xml2 by providing
a mechanism for supplying a prefix to a named namespace URI.
How does it support multiple namespaces in one document?
--
Peter Eisentraut
http://developer.postgresql.org/~petere/
On 3/6/07, Mike Rylander [EMAIL PROTECTED] wrote:
Attatched you'll find a patch that I've been kicking around for a
while that I'd like to propose for inclusion in 8.3. I attempted to
submit this through the original xml2 author (as far back as the 7.4
days) but got no response.
It's really
Tom Lane [EMAIL PROTECTED] wrote:
This is a stand-alone patch for aggressive freezing. I'll propose
to use OldestXmin instead of FreezeLimit as the freeze threshold
in the circumstances below:
I think it's a really bad idea to freeze that aggressively under any
circumstances except
On 3/6/07, Neil Conway [EMAIL PROTECTED] wrote:
On Tue, 2007-03-06 at 00:18 -0500, Tom Lane wrote:
Perhaps they provide a compatibility hack that you didn't spot?
Quite possibly. Anyone have any suggestions?
Python guys discuss the situation here:
http://www.python.org/dev/peps/pep-0353/
Dear PostgreSQL Contributors,
I seek help from designers, developers, testers,defect fixers,project
managers or playing any other key role in Free/Open Source software
development or maintenence
in carrying out a study on practices and problems of defect management in
various Free/Open Source
ITAGAKI Takahiro [EMAIL PROTECTED] writes:
I don't think we can supply such a historical database functionality here,
because we can guarantee it just only for INSERTed tuples even if we pay
attention. We've already enabled autovacuum as default, so that we cannot
predict when the next
On Tue, Mar 06, 2007 at 01:50:34PM +0530, sharath kumar wrote:
I need some help in building new index structure using GIST.
I have to include a procedure for restructuring the tree immediately after
i have created the index using gist. How do i do it?
Actually I am trying to implement the
Hi,
This is to get feedback to meet the following TODO:
- Simplify ability to create partitioned tables
This would allow creation of partitioned tables without requiring
creation of rules for INSERT/UPDATE/DELETE, and constraints for rapid
partition selection. Options could include
On Mon, 2007-03-05 at 22:25 +0530, Pavan Deolasee wrote:
Simon Riggs wrote:
On Mon, 2007-03-05 at 21:39 +0530, Pavan Deolasee wrote:
Currently each tuple is moved individually. You'd need to inspect the
whole HOT chain on a page, calculate space for that and then try to move
them all
Jim Nasby wrote:
On Mar 1, 2007, at 8:45 AM, Florian G. Pflug wrote:
Another possibility would be to move this setting into the
recovery.conf. The problems
with this approach is that the recovery.conf file is deleted after the
information
it contains is incorporated into pg_control. Thus, the
NikhilS wrote:
iv) Based on the PRIMARY, UNIQUE, REFERENCES information specified,
pass it on to the children tables.
How will you maintain a primary key in such a table, considering that
indexes can't span multiple tables?
1) Whether we should use triggers/rules for step number (iii) above.
Teodor Sigaev [EMAIL PROTECTED] writes:
A closer reading, however, shows that at least for cases like intarray,
btree_gist, etc., the detoasting of an index value is being done in the
gist decompress function, so the value seen via GISTENTRY in the other
functions should already have been
Hi,
On 3/6/07, Peter Eisentraut [EMAIL PROTECTED] wrote:
NikhilS wrote:
iv) Based on the PRIMARY, UNIQUE, REFERENCES information specified,
pass it on to the children tables.
How will you maintain a primary key in such a table, considering that
indexes can't span multiple tables?
We will
NikhilS [EMAIL PROTECTED] writes:
the intention is to use this information from the parent and make it a
property of the child table. This will avoid the step for the user having to
manually specify CREATE INDEX and the likes on all the children tables
one-by-one.
Missed the start of this
On 3/6/07, NikhilS [EMAIL PROTECTED] wrote:
Hi,
On 3/6/07, Gregory Stark [EMAIL PROTECTED] wrote:
NikhilS [EMAIL PROTECTED] writes:
the intention is to use this information from the parent and make it a
property of the child table. This will avoid the step for the user
having to
Hi,
On 3/6/07, Gregory Stark [EMAIL PROTECTED] wrote:
NikhilS [EMAIL PROTECTED] writes:
the intention is to use this information from the parent and make it a
property of the child table. This will avoid the step for the user having
to
manually specify CREATE INDEX and the likes on all the
On 2007-03-06, Gregory Stark [EMAIL PROTECTED] wrote:
Teodor Sigaev [EMAIL PROTECTED] writes:
A closer reading, however, shows that at least for cases like intarray,
btree_gist, etc., the detoasting of an index value is being done in the
gist decompress function, so the value seen via
ITAGAKI Takahiro [EMAIL PROTECTED] writes:
Tom Lane [EMAIL PROTECTED] wrote:
I think it's a really bad idea to freeze that aggressively under any
circumstances except being told to (ie, VACUUM FREEZE). When you
freeze, you lose history information that might be needed later --- for
forensic
The problem is that this is the only place in the code where we make wholesale
assumptions that a datum that comes from a tuple (heap tuple or index tuple)
isn't toasted. There are other places but they're always flagged with big
comments explaining *why* the datum can't be toasted and they're
I need some help in building new index structure using GIST.
I have to include a procedure for restructuring the tree immediately
after i have created the index using gist. How do i do it?
IMHO, you don't have a chance to do it without patching postgres or developing
your own low-level
On Mar 5, 2007, at 8:34 PM, Greg Smith wrote:
On Thu, 22 Feb 2007, Jim C. Nasby wrote:
It would also be extremely useful to make checkpoint stats visible
somewhere in the database (presumably via the existing stats
mechanism)... I'm thinking just tracking how many pages had to be
flushed
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi all,
I'm observing this behavior in 8.2:
kalman=# create table test ( a integer, b integer );
CREATE TABLE
kalman=#
kalman=# CREATE OR REPLACE FUNCTION sp_test ( INTEGER )
kalman-# RETURNS INTEGER AS'
kalman'# DECLARE
kalman'#a_id_contract
On Sun, 2007-03-04 at 10:14 -0800, Tom Lane wrote:
If you did this only once, the results are not really trustworthy;
you need to average several similar runs before you can have much
confidence. pgbench's inter-run variation is usually upwards of 10%,
so trying to draw conclusions about
I get it. scan.l converts '!=' to '':
644 /* Convert != operator to for
compatibility */
645 if (strcmp(yytext, !=) == 0)
646 yylval.str = pstrdup();
647 else
648
With copyout_max_cached being 8K:
^^^
4K
Working
set 16K 32K 64K 128K256K512K1M 2M 128M
Seconds 4.8
Andrew - Supernews [EMAIL PROTECTED] writes:
The places in the intarray code that you tried to fix in your patch at
the start of this thread are not dealing with data that came from a tuple,
but from data that came from a decompress method. It's expected that the
decompress method does the
On a 1P system system with 512K L2, it is more obvious why we shouldn't
bypass L2 for small reads:
The same readtest as my previous mail invoked as following:
./readtest -s working-set-size -f /platform/i86pc/boot_archive -n 100
With copyout_max_cached being 128K:
Working
set 16K
- Forwarded message from Sherry Moore [EMAIL PROTECTED] -
Date: Mon, 5 Mar 2007 21:34:19 -0800
From: Sherry Moore [EMAIL PROTECTED]
To: Tom Lane [EMAIL PROTECTED]
Cc: Luke Lonergan [EMAIL PROTECTED],
Mark Kirkwood [EMAIL PROTECTED],
Pavan Deolasee [EMAIL PROTECTED],
Hi Tom,
Sorry about the delay. I have been away from computers all day.
In the current Solaris release in development (Code name Nevada,
available for download at http://opensolaris.org), I have implemented
non-temporal access (NTA) which bypasses L2 for most writes, and reads
larger than
NikhilS wrote:
Hi,
On 3/6/07, Peter Eisentraut [EMAIL PROTECTED] wrote:
NikhilS wrote:
iv) Based on the PRIMARY, UNIQUE, REFERENCES information specified,
pass it on to the children tables.
How will you maintain a primary key in such a table, considering that
indexes can't span multiple
Teodor Sigaev [EMAIL PROTECTED] writes:
I'm afraid that we have some lack of understanding. Flow of operation with
indexed tuple in gist is:
- read tuple
- get n-th attribute with a help of index_getattr
- call user-defined decompress method which should, at least, detoast value
- result
On Mon, Mar 05, 2007 at 09:04:46PM -0600, Jim Nasby wrote:
FWIW, this sounds like a subset of the Query By Example stuff that
someone is working on. I don't have a URL handy since I'm on a plane,
but I think google can find it.
It's now called ObelisQ http://pgfoundry.org/projects/qbe
Gregory Stark [EMAIL PROTECTED] writes:
Andrew - Supernews [EMAIL PROTECTED] writes:
So I think you've mis-analyzed the problem. That's especially true since
you are claiming that the existing code is already buggy when in fact no
such bugs have been reported (and clearly intarray has been
JD,
Don't get me wrong, I want things to be easily understandable as well
but the reason you site above pretty much
makes us need to remove most of the postgresql.conf, including all
bgwriter, vacuum cost delay, and autovac settings.
Not to mention commit delay and others ;).
Wouldn't that
So when does index_form_tuple get called?
The single call of index_form_tuple in GiST core is in gistFormTuple which
initially compress any indexed value with a help of their compress methods.
Only tuples formed by gistFormTuple could be inserted in index.
Does every data type define a
On 3/6/07, Peter Eisentraut [EMAIL PROTECTED] wrote:
Mike Rylander wrote:
The patch adds support for default XML namespaces in xml2 by providing
a mechanism for supplying a prefix to a named namespace URI.
How does it support multiple namespaces in one document?
It supports one default
On 3/6/07, Nikolay Samokhvalov [EMAIL PROTECTED] wrote:
On 3/6/07, Mike Rylander [EMAIL PROTECTED] wrote:
Attatched you'll find a patch that I've been kicking around for a
while that I'd like to propose for inclusion in 8.3. I attempted to
submit this through the original xml2 author (as far
And it's not clear _int_gist.c has been running with toasted array values for
years because it's limited to arrays of 100 integers (or perhaps 200 integers,
there's a factor of 2 in the test). That's not enough to trigger toasting
unless there are other large columns in the same table.
That's was
Does every data type define a compress/decompress method? Even if it's not a
data type that normally gets very large?
Yes, any GiST opclass should have such methods. In trivial case it just
returns
input value. As I remember, only R-Tree emulation over boxes, contrib/seg and
contrib/cube
[EMAIL PROTECTED] (Jim Nasby) writes:
FWIW, this sounds like a subset of the Query By Example stuff that
someone is working on. I don't have a URL handy since I'm on a plane,
but I think google can find it.
The pgFoundry project is here...
http://pgfoundry.org/projects/qbe
And yes, indeed,
[EMAIL PROTECTED] (Peter Eisentraut) writes:
NikhilS wrote:
iv) Based on the PRIMARY, UNIQUE, REFERENCES information specified,
pass it on to the children tables.
How will you maintain a primary key in such a table, considering that
indexes can't span multiple tables?
On the one hand, I
input value. As I remember, only R-Tree emulation over boxes, contrib/seg and
contrib/cube have simple compress method.
Hm, if they just return the original datum without detoasting it then it could
be an issue. I'll check.
seg and box aren't a varlena types, but cube is and it seems broken :(.
[ cc'd to pgsql-jdbc which seems the group most likely to be affected
by any protocol change ]
So I've been working on a plan cache module per my earlier proposal,
and I've run up against a problem with getting exec_parse_message
to use it. The problem is that the current rather hackish
And yes, indeed, this sounds quite a lot like what Meredith Patterson
presented at the Toronto conference.
This would be good to have, though, since Meredith's work has some problematic
IP encumbrances.
Question, though: is the SKYLINE syntax part of a standard anywhere?
--
Josh Berkus
On Wed, Mar 07, 2007 at 02:43:39AM +1030, Shane Ambler wrote:
I think a way can be devised to maintain the primary key and unique
constraints.
If a search is done on the parent table, the planner knows to rewrite
the query as a union (or union all) of all child tables that relate to
the
Gaetano Mendola [EMAIL PROTECTED] writes:
[ 8.2 evaluates volatile functions in the targetlist of a view ]
If I mark the function as STABLE or IMMUTABLE then even with version
8.2 the function is not evaluated. Is this the intended behavior?
Yes; people complained that we needed to be more
Teodor Sigaev [EMAIL PROTECTED] writes:
And it's not clear _int_gist.c has been running with toasted array values for
years because it's limited to arrays of 100 integers (or perhaps 200
integers,
there's a factor of 2 in the test). That's not enough to trigger toasting
unless there are
On Mon, 2007-03-05 at 21:02 -0700, Jim Nasby wrote:
On Mar 5, 2007, at 2:03 PM, Heikki Linnakangas wrote:
Another approach I proposed back in December is to not have a
variable like that at all, but scan the buffer cache for pages
belonging to the table you're scanning to initialize the
Jeff Davis [EMAIL PROTECTED] writes:
If I were to implement this idea, I think Heikki's bitmap of pages
already read is the way to go.
I think that's a good way to guarantee that you'll not finish in time
for 8.3. Heikki's idea is just at the handwaving stage at this point,
and I'm not even
Josh Berkus josh@agliodbs.com writes:
Question, though: is the SKYLINE syntax part of a standard anywhere?
There's certainly not anything like that in SQL2003.
I'm also kind of wondering if the main use-cases couldn't be met with
suitable multi-input custom aggregates, which is something we
Tom Lane [EMAIL PROTECTED] writes:
B. Don't store the unnamed statement in the plan cache. To make sure
it's not used anymore when the plan might be stale, forcibly discard
the unnamed statement after execution. This would get rid of a lot
of overhead but would mean a significant change in
On Tue, 2007-03-06 at 12:59 -0500, Tom Lane wrote:
Jeff Davis [EMAIL PROTECTED] writes:
If I were to implement this idea, I think Heikki's bitmap of pages
already read is the way to go.
I think that's a good way to guarantee that you'll not finish in time
for 8.3. Heikki's idea is just
Gregory Stark [EMAIL PROTECTED] writes:
Can we forcibly discard it if *any* messages are received that might
invalidate a plan? So basically it would work fine unless anyone in the system
does any DDL at all? I guess that has the downside of introducing random
unpredictable failures.
Ugh :-(
Martijn van Oosterhout wrote:
On Wed, Mar 07, 2007 at 02:43:39AM +1030, Shane Ambler wrote:
I think a way can be devised to maintain the primary key and unique
constraints.
If a search is done on the parent table, the planner knows to rewrite
the query as a union (or union all) of all child
Jeff Davis wrote:
On Mon, 2007-03-05 at 21:02 -0700, Jim Nasby wrote:
On Mar 5, 2007, at 2:03 PM, Heikki Linnakangas wrote:
Another approach I proposed back in December is to not have a
variable like that at all, but scan the buffer cache for pages
belonging to the table you're scanning to
NikhilS wrote:
We will not (I know its a hard thing to do :) ), the intention is to
use this information from the parent and make it a property of the
child table. This will avoid the step for the user having to manually
specify CREATE INDEX and the likes on all the children tables
Florian,
This sounds like what is really needed is a way to lock a certain
condition, namely the existance or non-existance of a record with
certain values in certain fields. This would not only help this case,
it would also help RI triggers, because those wouldn't have to acquire
a share
Tom,
I'm also kind of wondering if the main use-cases couldn't be met with
suitable multi-input custom aggregates, which is something we already
have as of 8.2.
Actually, given that skyline of is *only* for aggregate sorting (as far as I
can tell) it doesn't present the complications which
Peter Eisentraut [EMAIL PROTECTED] writes:
But when I say
CREATE TABLE ( a int PRIMARY KEY, ... ) PARTITION blah ...
then I expect that the primary key will be enforced across all
partitions. We currently sidestep that issue by not offering seemingly
transparent partitioning. But if you
I think C is how the JDBC driver is written. We name the statements
if they have been used more than prepareThreshold times.
So we have a mechanism by which to allow statements to be cached, or
not.
Dave
On 6-Mar-07, at 1:14 PM, Tom Lane wrote:
Gregory Stark [EMAIL PROTECTED] writes:
It doesn't actually crash, it just fails CHECKARRVALID. I added an assertion
in there to cause it to generate a core dump.
Wow, catch that, see attached patch.
g_int_decompress doesn't returns detoasted array in case it was empty.
Previously it was safe because empty array never has been
Tom Lane wrote:
Jeff Davis [EMAIL PROTECTED] writes:
If I were to implement this idea, I think Heikki's bitmap of pages
already read is the way to go.
I think that's a good way to guarantee that you'll not finish in time
for 8.3. Heikki's idea is just at the handwaving stage at this point,
Agreed, it needs to Just Work. I think it'd still be useful though
if we only support auto-partitioning on the primary key, and that
restriction avoids the indexing problem.
+1
--
Josh Berkus
PostgreSQL @ Sun
San Francisco
---(end of
Tom Lane wrote:
Gregory Stark [EMAIL PROTECTED] writes:
Can we forcibly discard it if *any* messages are received that might
invalidate a plan? So basically it would work fine unless anyone in the system
does any DDL at all? I guess that has the downside of introducing random
unpredictable
Josh Berkus wrote:
Tom,
I'm also kind of wondering if the main use-cases couldn't be met with
suitable multi-input custom aggregates, which is something we already
have as of 8.2.
Actually, given that skyline of is *only* for aggregate sorting (as far as
I
can tell) it doesn't
Tom Lane wrote:
Peter Eisentraut [EMAIL PROTECTED] writes:
But when I say
CREATE TABLE ( a int PRIMARY KEY, ... ) PARTITION blah ...
then I expect that the primary key will be enforced across all
partitions. We currently sidestep that issue by not offering seemingly
transparent
On Fri, Mar 02, 2007 at 06:59:50PM -0500, Tom Lane wrote:
David Fetter [EMAIL PROTECTED] writes:
1. In src/backend/commands/tablecmds.c, change DefineRelation as
follows:
* After the first call to heap_create_with_catalog, construct and
do another call to for the array type.
David Fetter [EMAIL PROTECTED] writes:
I noticed something in src/backend/commands/tablecmds.c which worries
me, namely that it ignores functions and views.
What?
regards, tom lane
---(end of broadcast)---
TIP 3: Have you
On Tue, Mar 06, 2007 at 04:14:07PM -0500, Tom Lane wrote:
David Fetter [EMAIL PROTECTED] writes:
I noticed something in src/backend/commands/tablecmds.c which
worries me, namely that it ignores functions and views.
What?
The it in question is, find_composite_type_dependencies()
Cheers,
D
David Fetter [EMAIL PROTECTED] writes:
On Tue, Mar 06, 2007 at 04:14:07PM -0500, Tom Lane wrote:
David Fetter [EMAIL PROTECTED] writes:
I noticed something in src/backend/commands/tablecmds.c which
worries me, namely that it ignores functions and views.
What?
The it in question is,
On Tue, Mar 06, 2007 at 04:24:36PM -0500, Tom Lane wrote:
David Fetter [EMAIL PROTECTED] writes:
On Tue, Mar 06, 2007 at 04:14:07PM -0500, Tom Lane wrote:
David Fetter [EMAIL PROTECTED] writes:
I noticed something in src/backend/commands/tablecmds.c which
worries me, namely that it
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Josh Berkus
Sent: dinsdag 6 maart 2007 19:45
To: pgsql-hackers@postgresql.org
Cc: Florian G. Pflug; Martijn van Oosterhout; Shane Ambler;
NikhilS; Peter Eisentraut
Subject: Re: [HACKERS] Auto creation of
On Mon, 2007-03-05 at 21:34 -0800, Sherry Moore wrote:
- Based on a lot of the benchmarks and workloads I traced, the
target buffer of read operations are typically accessed again
shortly after the read, while writes are usually not. Therefore,
the default operation
On Tue, 2007-03-06 at 12:22 -0500, Tom Lane wrote:
A. Just accept the extra overhead, thereby preserving the current
behavior of unnamed statements, and gaining the benefit that plan
invalidation will work correctly in the few cases where an unnamed
statement's plan lasts long enough to need
Simon Riggs [EMAIL PROTECTED] writes:
On Tue, 2007-03-06 at 12:22 -0500, Tom Lane wrote:
A. Just accept the extra overhead, thereby preserving the current
behavior of unnamed statements, and gaining the benefit that plan
invalidation will work correctly in the few cases where an unnamed
Teodor Sigaev [EMAIL PROTECTED] writes:
It doesn't actually crash, it just fails CHECKARRVALID. I added an assertion
in there to cause it to generate a core dump.
Wow, catch that, see attached patch.
g_int_decompress doesn't returns detoasted array in case it was empty.
Previously it was
On Tue, 2007-03-06 at 18:47 +, Heikki Linnakangas wrote:
Tom Lane wrote:
Jeff Davis [EMAIL PROTECTED] writes:
If I were to implement this idea, I think Heikki's bitmap of pages
already read is the way to go.
I think that's a good way to guarantee that you'll not finish in time
On Tue, 6 Mar 2007, Alvaro Herrera wrote:
Also, keep in mind that there were plenty of changes in the executor.
This stuff is not likely to be very easy to implement efficiently using
our extant executor machinery; note that Ranbeer mentioned
implementation of block nested loop and other
On Mar 6, 2007, at 12:17 AM, Tom Lane wrote:
Jim Nasby [EMAIL PROTECTED] writes:
An idea I've been thinking about would be to have the bgwriter or
some other background process actually try and keep the free list
populated,
The bgwriter already tries to keep pages just in front of the clock
On Mar 6, 2007, at 10:56 AM, Jeff Davis wrote:
We also don't need an exact count, either. Perhaps there's some way
we could keep a counter or something...
Exact count of what? The pages already in cache?
Yes. The idea being if you see there's 10k pages in cache, you can
likely start 9k
Gregory Stark [EMAIL PROTECTED] wrote:
The hoped for gain here is that vacuum finds fewer pages with tuples that
exceed vacuum_freeze_min_age? That seems useful though vacuum is still going
to have to read every page and I suspect most of the writes pertain to dead
tuples, not freezing
On Tue, 2007-03-06 at 17:43 -0700, Jim Nasby wrote:
On Mar 6, 2007, at 10:56 AM, Jeff Davis wrote:
We also don't need an exact count, either. Perhaps there's some way
we could keep a counter or something...
Exact count of what? The pages already in cache?
Yes. The idea being if you see
Tom Lane [EMAIL PROTECTED] wrote:
I said nothing about expired tuples. The point of not freezing is to
preserve information about the insertion time of live tuples.
I don't know what good it will do -- for debugging?
Why don't you use CURRENT_TIMESTAMP?
And your
test case is
On Tue, 2007-03-06 at 18:29 +, Heikki Linnakangas wrote:
Jeff Davis wrote:
On Mon, 2007-03-05 at 21:02 -0700, Jim Nasby wrote:
On Mar 5, 2007, at 2:03 PM, Heikki Linnakangas wrote:
Another approach I proposed back in December is to not have a
variable like that at all, but scan the
Incidentally, we tried triggering NTA (L2 cache bypass) unconditionally and in
various patterns and did not see the substantial gain as with reducing the
working set size.
My conclusion: Fixing the OS is not sufficient to alleviate the issue. We see
a 2x penalty (1700MB/s versus 3500MB/s) at
On Tue, 6 Mar 2007, Jim Nasby wrote:
The flipside is that it's much easier to machine-parse a table rather
than trying to scrape the logs.
Now you might realize why I've been so vocal on the SQL log export
implementation details.
And I don't think we'll generally care about each individual
ITAGAKI Takahiro [EMAIL PROTECTED] writes:
Tom Lane [EMAIL PROTECTED] wrote:
I said nothing about expired tuples. The point of not freezing is to
preserve information about the insertion time of live tuples.
I don't know what good it will do -- for debugging?
Exactly. As an example, I've
Greg Smith [EMAIL PROTECTED] wrote:
After a few months of staring at this data, I've found averages like that
misleading. The real problem areas correlate with the peak pages written
at any one checkpoint. Lowering that value is really the end-game for
optimizing the background writer,
Hi Shane,
Maybe I'm looking at auto-maintenance which is beyond any current planning?
Many of your suggestions are useful, but auto-maintenance will be beyond the
current plan.
Regards,
Nikhils
EnterpriseDB http://www.enterprisedb.com
Hi,
On 3/7/07, Tom Lane [EMAIL PROTECTED] wrote:
Peter Eisentraut [EMAIL PROTECTED] writes:
But when I say
CREATE TABLE ( a int PRIMARY KEY, ... ) PARTITION blah ...
then I expect that the primary key will be enforced across all
partitions. We currently sidestep that issue by not offering
Andrew Dunstan wrote:
Tom Lane wrote:
Peter Eisentraut [EMAIL PROTECTED] writes:
But when I say
CREATE TABLE ( a int PRIMARY KEY, ... ) PARTITION blah ...
then I expect that the primary key will be enforced across all
partitions. We currently sidestep that issue by not offering
seemingly
Ühel kenal päeval, T, 2007-03-06 kell 18:28, kirjutas Jeff Davis:
On Tue, 2007-03-06 at 18:29 +, Heikki Linnakangas wrote:
Jeff Davis wrote:
On Mon, 2007-03-05 at 21:02 -0700, Jim Nasby wrote:
On Mar 5, 2007, at 2:03 PM, Heikki Linnakangas wrote:
Another approach I proposed back in
Maybe. The most obvious use for automatic partitioning that I can
think of would be based in the value of a timestamptz field rather
than any PK. Of course I tend to work more in the OLTP field than in
DW type apps, where other considerations might apply.
I second that - partitioning on
96 matches
Mail list logo