Re: [HACKERS] [PATCHES] update i386 spinlock for hyperthreading

2004-02-28 Thread Kenneth Marshall
On Fri, Feb 20, 2004 at 05:26:46AM -0500, Neil Conway wrote: Simon Riggs [EMAIL PROTECTED] writes: Kenneth Marshall would like me to post this: I agree that in order to manage today's large memory machines, we need to have less contention in our buffer management strategies. The two main

Re: [HACKERS] Why are these ARC variables per-backend?

2004-04-20 Thread Kenneth Marshall
On Mon, Apr 19, 2004 at 02:58:11PM -0400, Tom Lane wrote: Jan Wieck [EMAIL PROTECTED] writes: Tom Lane wrote: I've got a problem with these variables in freelist.c: static int strategy_cdb_found; static int strategy_cdb_replace; These two most definitely are per

Re: [HACKERS] ARC Memory Usage analysis

2004-10-24 Thread Kenneth Marshall
On Fri, Oct 22, 2004 at 03:35:49PM -0400, Jan Wieck wrote: On 10/22/2004 2:50 PM, Simon Riggs wrote: I've been using the ARC debug options to analyse memory usage on the PostgreSQL 8.0 server. This is a precursor to more complex performance analysis work on the OSDL test suite. I've

Re: [PATCHES] [HACKERS] ARC Memory Usage analysis

2004-10-27 Thread Kenneth Marshall
On Mon, Oct 25, 2004 at 05:53:25PM -0400, Tom Lane wrote: Greg Stark [EMAIL PROTECTED] writes: So I would suggest using something like 100us as the threshold for determining whether a buffer fetch came from cache. I see no reason to hardwire such a number. On any hardware, the

Re: [HACKERS] Minor TODO list changes

2004-11-07 Thread Kenneth Marshall
Bruce, Just to chime in. I also agree that fillfactor is useful. I have been investigating different index variants and different fill factors can greatly influence the performance of the index. I also think it may play a key role in minimizing the small table/ many inserts/updates performance

Re: [Testperf-general] Re: [HACKERS] ExclusiveLock

2004-11-23 Thread Kenneth Marshall
On Tue, Nov 23, 2004 at 12:04:17AM +, Simon Riggs wrote: On Mon, 2004-11-22 at 23:37, Greg Stark wrote: Simon Riggs [EMAIL PROTECTED] writes: - Find a way to reduce rotational delay when repeatedly writing last WAL page Currently fsync of WAL requires the disk platter to

[HACKERS] Solaris 8 regression test failure with 8.0.0beta5

2004-11-24 Thread Kenneth Marshall
Here are the diffs for the regression test failures on Solaris 8. The tests work fine on Redhat9 and Redhat Enterprise Linux 3. Ken Marshall *** ./expected/errors.out Sat Mar 13 22:25:17 2004 --- ./results/errors.outTue Nov 23 14:09:45 2004 *** *** 297,303 --

Re: [Testperf-general] Re: [HACKERS] ExclusiveLock

2004-11-24 Thread Kenneth Marshall
On Wed, Nov 24, 2004 at 11:00:30AM -0500, Bort, Paul wrote: From: Kenneth Marshall [mailto:[EMAIL PROTECTED] [snip] The simplest idea I had was to pre-layout the WAL logs in a contiguous fashion on the disk. Solaris has this ability given appropriate FS parameters and we should

[HACKERS] follow-up to previous build problem for 8.0.0beta5 on SPARC

2004-11-24 Thread Kenneth Marshall
The failure that I posted earlier for 8.0.0beta5 on Solaris 8/SPARC with gcc-3.4.0 and -O3 can be worked around by disabling the interblock scheduling. I used the following gcc options and 8.0.0beta5 built fine on the SPARC Solaris 8 machine: gcc -O3 -fno-sched-interblock ... The Redhat 9 and

Re: [HACKERS] commit_delay, siblings

2005-06-29 Thread Kenneth Marshall
On Wed, Jun 29, 2005 at 08:14:36AM +0100, Simon Riggs wrote: Group commit is a well-documented technique for improving performance, but the gains only show themselves on very busy systems. It is possible in earlier testing any apparent value was actually hidden by the BufMgrLock issues we

Re: [HACKERS] 2PC transaction id

2005-07-02 Thread Kenneth Marshall
It certainly helps if you need to debug a process. Ken On Fri, Jul 01, 2005 at 09:06:03PM +0300, Heikki Linnakangas wrote: On Fri, 1 Jul 2005, Oliver Jowett wrote: PS: noticed in passing: psql's help doesn't seem to know about the 2PC command syntax yet. True. Should we add support

Re: [HACKERS] Checkpoint cost, looks like it is WAL/CRC

2005-07-07 Thread Kenneth Marshall
On Thu, Jul 07, 2005 at 11:36:40AM -0400, Tom Lane wrote: Greg Stark [EMAIL PROTECTED] writes: Tom Lane [EMAIL PROTECTED] writes: What we *could* do is calculate a page-level CRC and store it in the page header just before writing out. Torn pages would then manifest as a wrong CRC on

Re: [pgsql-advocacy] [HACKERS] MySQL to PostgreSQL for SugarCRM

2005-08-09 Thread Kenneth Marshall
Compiere also runs on PostgreSQL and has gotten some good press. It previously only support Oracle. Ken On Tue, Aug 02, 2005 at 08:53:52PM +0300, Hannu Krosing wrote: On L, 2005-07-30 at 22:26 -0400, Denis Lussier wrote: Thanks, I'll check it out. I didn't see much evidence on the SugarCRM

Re: [HACKERS] [ANNOUNCE] PostgreSQL 8.1 Beta 4

2005-10-30 Thread Kenneth Marshall
Marc, I just finished a build with the 8.1beta4 for Solaris 8 (SPARC) with: OpenSSL 0.9.8 Heimdal 0.7 tcl/tk-8.4.8 perl-5.8.4 using gcc-3.4.3 and the following compile command: gcc -O3 -DOPENSSL_DISABLE_OLD_DES_SUPPORT -fno-sched-interblock All tests passed successfully. Ken On Mon, Oct

Re: [HACKERS] [ANNOUNCE] PostgreSQL 8.1 RC1

2005-11-06 Thread Kenneth Marshall
Marc, Okay, I found an OpenSSL-0.9.7 and readline library. The IRIX 6.5 IP35 also passed with the OpenSSL and readline included. This is with the IRIX cc and not gcc. Ken On Mon, Oct 24, 2005 at 11:51:26AM -0300, Marc G. Fournier wrote: We have released a Release Candidate 1 of the

Re: [HACKERS] [ANNOUNCE] PostgreSQL 8.1 Beta 4

2005-11-06 Thread Kenneth Marshall
Marc, I just finished a build with the 8.1beta4 for IRIX 6.5 but without the nuances. We do not really use SGI other than in special circumstances but the regression test passed all tests: configure --without-readline using IRIX cc. Ken On Mon, Oct 24, 2005 at 11:51:26AM -0300, Marc G.

Re: Buildfarm coverage (was Re: [HACKERS] OK, ready for RC1 or Beta6)

2004-12-03 Thread Kenneth Marshall
On Fri, Dec 03, 2004 at 03:20:48PM -0500, Tom Lane wrote: Joshua D. Drake [EMAIL PROTECTED] writes: 1. Buildfarm doesn't yet have that many platforms on it. It's not as bad as all that. Our current list of supported platforms (ie, things that got tested last time) is AIX

Re: [HACKERS] V8 Beta 5 on AIX

2004-12-06 Thread Kenneth Marshall
On Mon, Dec 06, 2004 at 12:53:52PM -0500, Bruce Momjian wrote: Brad Nicholson wrote: OK, I assume you used --enable-thread-safety in configure. Correct. This should have added some PTHREAD link flags to your libpq build, and those settings should have followed the libpq library

Re: [HACKERS] Call for port reports

2004-12-11 Thread Kenneth Marshall
Port report for Solaris 8: No errors. uname -a: SunOS sunos58.build 5.8 Generic_117350-11 sun4u sparc SUNW,UltraAX-i2 gcc -v: Reading specs from /gcc-3.4.0/sunos5/bin/../lib/gcc/sparc-sun-solaris2.8/3.4.0/specs Configured with: /gcc-3.4.0/src/dist/configure --prefix=/usr/site/gcc-3.4.0

Re: [HACKERS] RC2 and open issues

2004-12-24 Thread Kenneth Marshall
On Mon, Dec 20, 2004 at 11:20:46PM -0500, Tom Lane wrote: Bruce Momjian pgman@candle.pha.pa.us writes: Tom Lane wrote: Exactly. But 1% would be uselessly small with this definition. Offhand I'd think something like 50% might be a starting point; maybe even more. What that says is that a

Re: [HACKERS] Two-phase commit for 8.1

2005-01-20 Thread Kenneth Marshall
On Wed, Jan 19, 2005 at 07:42:03PM -0500, Tom Lane wrote: Marc G. Fournier [EMAIL PROTECTED] writes: If the patch is ready to be committed early in the cycle, I'd say most definitely ... just depends on how late in the cycle its ready ... My recollection is that it's quite far from being

Re: [HACKERS] ARC patent

2005-01-25 Thread Kenneth Marshall
On Fri, Jan 21, 2005 at 03:42:38PM +0100, Manfred Koizar wrote: On Fri, 21 Jan 2005 02:31:40 +0200, Hannu Krosing [EMAIL PROTECTED] wrote: 2) Another simple, but nondeterministic, hack would be using randomness, i.e. 2.1) select a random buffer in LR side half (or 30% or 60%) of

Re: [HACKERS] LWLock cache line alignment

2005-02-03 Thread Kenneth Marshall
On Thu, Feb 03, 2005 at 06:26:16AM -0800, Simon Riggs wrote: From: Tom Lane [mailto:[EMAIL PROTECTED] wrote Simon Riggs [EMAIL PROTECTED] writes: It looks like padding out LWLock struct would ensure that each of those were in separate cache lines? I've looked at this before and I

Re: [HACKERS] Thinking about breaking up the BufMgrLock

2005-02-08 Thread Kenneth Marshall
On Sun, Feb 06, 2005 at 07:30:37PM -0500, Tom Lane wrote: ReadBuffer needs to do a lookup to map the page ID to a buffer ID, which in principle requires only a shared lock on the page-to-buffer mapping (embodied in the buf_table hash table). Assuming success, it also needs to mark the

Re: [HACKERS] Design notes for BufMgrLock rewrite

2005-02-16 Thread Kenneth Marshall
On Wed, Feb 16, 2005 at 12:33:38PM -0500, Tom Lane wrote: Jim C. Nasby [EMAIL PROTECTED] writes: The advantage of using a counter instead of a simple active bit is that buffers that are (or have been) used heavily will be able to go through several sweeps of the clock before being freed.

Re: [HACKERS] left-deep plans?

2005-02-22 Thread Kenneth Marshall
On Tue, Feb 22, 2005 at 05:40:40PM +1100, Neil Conway wrote: Tom Lane wrote: Yes, and it's been rejected. The notion is obviously bogus; it amounts to assuming that every database is a star schema with only one core table. Interesting; yes, I suppose that's true. Once we get into GEQO

Re: [HACKERS] left-deep plans?

2005-02-23 Thread Kenneth Marshall
On Wed, Feb 23, 2005 at 10:02:22AM +1100, Neil Conway wrote: Kenneth Marshall wrote: GEQO is an attempt to provide a near-optimal join order without using an exhaustive search. An exhaustive, deterministic search of a subset of the search space has a non-zero probability of finding only

Re: [HACKERS] Changing the default wal_sync_method to open_sync for Win32?

2005-03-20 Thread Kenneth Marshall
On Wed, Mar 16, 2005 at 11:20:12PM -0500, Bruce Momjian wrote: Basically we do open_datasync - fdatasync - fsync. This is empirically what we found to be fastest on most operating systems, and we default to the first one that exists on the operating system. Notice we never default to

Re: [HACKERS] State of Kerberos v4 support

2005-05-06 Thread Kenneth Marshall
On Fri, May 06, 2005 at 05:00:36PM +0200, Magnus Hagander wrote: I'm working over the kerberos service principal name patch that's in the queue to make it good enough for application. During which I noticed it touches both kerberos 4 and kerberos 5 code, which leads me to two questions: 1)

Re: [HACKERS] proposed TODO: non-locking CREATE INDEX / REINDEX

2005-06-12 Thread Kenneth Marshall
On Fri, Jun 10, 2005 at 12:12:05PM -0400, Tom Lane wrote: Hannu Krosing [EMAIL PROTECTED] writes: As the number of tuples between CTID_INDEX_MIN and CTID_INDEX_MAX is finite, they must be added in finite time, by which time the index will be up-to-date and usable for querie planner. (i.e.

Re: [HACKERS] Shared locking in slru.c

2005-12-01 Thread Kenneth Marshall
On Wed, Nov 30, 2005 at 03:23:55PM -0500, Tom Lane wrote: Kenneth Marshall [EMAIL PROTECTED] writes: ... In pseudo-code, the operations to read the control information are: WriteControl: 1. Set latch. 2. Update control information 3. Increment latch version number. 4. Unset latch

Re: [HACKERS] Shared locking in slru.c

2005-12-01 Thread Kenneth Marshall
On Wed, Nov 30, 2005 at 01:53:13PM -0500, Tom Lane wrote: I've been looking at various ways to resolve this, but one thing that seems promising is to hack slru.c to take the control lock in shared mode, not exclusive mode, for read-only accesses to pages that are already in memory. The vast

Re: [HACKERS] Warm-cache prefetching

2005-12-09 Thread Kenneth Marshall
On Fri, Dec 09, 2005 at 10:37:25AM -0500, Bruce Momjian wrote: Kenneth Marshall wrote: The main benefit of pre-fetching optimization is to allow just- in-time data delivery to the processor. There are numerous papers illustrating the dramatic increase in data throughput by using

Re: [HACKERS] Warm-cache prefetching

2005-12-09 Thread Kenneth Marshall
The main benefit of pre-fetching optimization is to allow just- in-time data delivery to the processor. There are numerous papers illustrating the dramatic increase in data throughput by using datastructures designed to take advantage of prefetching. Factors of 3-7 can be realized and this can

Re: [HACKERS] Warm-cache prefetching

2005-12-10 Thread Kenneth Marshall
On Fri, Dec 09, 2005 at 11:32:48AM -0500, Bruce Momjian wrote: Tom Lane wrote: Bruce Momjian pgman@candle.pha.pa.us writes: I can see that being useful for a single-user application that doesn't have locking or I/O bottlenecks, and doesn't have a multi-stage design like a database. Do

Re: [HACKERS] Improving N-Distinct estimation by ANALYZE

2006-01-08 Thread Kenneth Marshall
On Fri, Jan 06, 2006 at 06:36:52PM -0500, Greg Stark wrote: Josh Berkus josh@agliodbs.com writes: These numbers don't make much sense to me. It seems like 5% is about as slow as reading the whole file which is even worse than I expected. I thought I was being a bit pessimistic to think

Re: [HACKERS] SpeedComparison

2006-02-12 Thread Kenneth Marshall
On Sat, Feb 11, 2006 at 09:21:43PM +0100, Jochem van Dieten wrote: On 2/11/06, Andrej Ricnik-Bay wrote: Has anyone here seen this one before? Do the values appear realistic? http://www.sqlite.org/cvstrac/wiki?p=SpeedComparison The values appear to originate from an intrsinsically

Re: [HACKERS] More thoughts about planner's cost estimates

2006-06-02 Thread Kenneth Marshall
Josh, Greg, and Tom, I do not know how sensitive the plans will be to the correlation, but one thought might be to map the histogram X histogram correlation to a square grid of values. Then you can map them to an integer which would give you 8 x 8 with binary values, a 5 x 5 with 4 values per

Re: [HACKERS] Duplicate rows sneaking in despite PRIMARY KEY / UNIQUE constraint

2006-06-07 Thread Kenneth Marshall
Travis, We have used postgresql 7.4, 8.0, and 8.1 with DSPAM and have never had a single problem like you are describing. In the past on this mailing list, these sorts of issues have been caused by hardware problems on the DB server in some cases. Good luck with tracking it down. Ken On Tue,

Re: [HACKERS] Rethinking stats communication mechanisms

2006-06-22 Thread Kenneth Marshall
On Sun, Jun 18, 2006 at 05:26:16PM -0400, Tom Lane wrote: I wrote: PFC [EMAIL PROTECTED] writes: So, the proposal : On executing a command, Backend stores the command string, then overwrites the counter with (counter + 1) and with the timestamp of command start. Periodically, like

Re: [HACKERS] Rethinking stats communication mechanisms

2006-06-22 Thread Kenneth Marshall
On Sun, Jun 18, 2006 at 11:07:41PM -0400, Tom Lane wrote: Bort, Paul [EMAIL PROTECTED] writes: Anyone know a variant of this that really works? Here's a theory: If the counter is bumped to an odd number before modification, and an even number after it's done, then the reader will know

Re: [HACKERS] A couple thoughts about btree fillfactor

2006-07-11 Thread Kenneth Marshall
On Mon, Jul 10, 2006 at 12:36:34PM -0400, Tom Lane wrote: Now that the index options infrastructure is in, I am having a couple of second thoughts about the specific behavior that's been implemented, particularly for btree fillfactor. 1. ... I'm thinking we could change the nbtsort.c code

Re: [HACKERS] Hash indexes (was: On-disk bitmap index patch)

2006-07-30 Thread Kenneth Marshall
On Fri, Jul 28, 2006 at 12:14:49PM -0500, Jim C. Nasby wrote: On Thu, Jul 27, 2006 at 01:46:01PM -0400, Alvaro Herrera wrote: Jim Nasby wrote: On Jul 25, 2006, at 3:31 PM, Tom Lane wrote: Hannu Krosing [EMAIL PROTECTED] writes: What would be the use-case for hash indexes ? And what

Re: [HACKERS] Hash indexes (was: On-disk bitmap index patch)

2006-08-03 Thread Kenneth Marshall
On Tue, Aug 01, 2006 at 02:26:18PM -0700, [EMAIL PROTECTED] wrote: Kenneth Marshall wrote: On Fri, Jul 28, 2006 at 12:14:49PM -0500, Jim C. Nasby wrote: On Thu, Jul 27, 2006 at 01:46:01PM -0400, Alvaro Herrera wrote: Jim Nasby wrote: On Jul 25, 2006, at 3:31 PM, Tom Lane wrote

Re: [HACKERS] 8.2 features status

2006-08-06 Thread Kenneth Marshall
On Fri, Aug 04, 2006 at 12:40:36PM -0400, Tom Lane wrote: Guillaume Smet [EMAIL PROTECTED] writes: And what about compression of on-disk sorting? That's purely a performance issue, which some people seem to want to define as not a new feature ... which is not *my* view of what's important

Re: BugTracker (Was: Re: [HACKERS] 8.2 features status)

2006-08-15 Thread Kenneth Marshall
RT is easy to setup/configure/use and works well with PostgreSQL as the backend. CPAN uses it for their bug tracker. Was there a list of features and requirements? Ken On Tue, Aug 15, 2006 at 10:59:52AM -0300, Marc G. Fournier wrote: On Fri, 11 Aug 2006, Alvaro Herrera wrote: I am suggesting

Re: BugTracker (Was: Re: [HACKERS] 8.2 features status)

2006-08-17 Thread Kenneth Marshall
On Wed, Aug 16, 2006 at 06:52:21AM +0200, Peter Eisentraut wrote: Tom Lane wrote: that the bug tracker would have to have a reasonable output email capability, but I'd not necessarily insist on being able to input to it by mail. Red Hat's present bugzilla system could be described that

Re: BugTracker (Was: Re: [HACKERS] 8.2 features status)

2006-08-17 Thread Kenneth Marshall
On Wed, Aug 16, 2006 at 01:22:43PM +0900, Michael Glaesemann wrote: On Aug 16, 2006, at 12:29 , Tom Lane wrote: So my current take on this would be that the bug tracker would have to have a reasonable output email capability, but I'd not necessarily insist on being able to input to it by

Re: BugTracker (Was: Re: [HACKERS] 8.2 features status)

2006-08-17 Thread Kenneth Marshall
, Kenneth Marshall wrote: RT is easy to setup/configure/use and works well with PostgreSQL as the backend. CPAN uses it for their bug tracker. Was there a list of features and requirements? I don't know if we ever came up with one, but I know that the big deal killer for a bug tracker

Re: [HACKERS] horo(r)logy test fail on solaris (again and solved)

2006-09-27 Thread Kenneth Marshall
On Wed, Sep 27, 2006 at 04:09:18PM +0200, Zdenek Kotala wrote: Tom Lane napsal(a): Zdenek Kotala [EMAIL PROTECTED] writes: The problem was generated, because -fast option was set only for the compiler and not for the linker. Linker takes wrong version of libraries. If -fast is set for

Re: [HACKERS] Need more reviewers!

2008-09-04 Thread Kenneth Marshall
On Thu, Sep 04, 2008 at 02:01:18PM -0400, Tom Lane wrote: Jonah H. Harris [EMAIL PROTECTED] writes: I'll push forward on reviewing and testing Xiao's hash index improvements for inclusion into core. Though, someone will still need to review my stuff. I think what the hash index patch

Re: [HACKERS] 8.3 vs HEAD difference in Interval output?

2008-10-09 Thread Kenneth Marshall
On Thu, Oct 09, 2008 at 11:50:17AM -0700, Ron Mayer wrote: Kevin Grittner wrote: Kevin Grittner [EMAIL PROTECTED] wrote: Even more surprising is the behavior for interval(1) here: [ some context with nonsurprising examples removed ...] ccdev=# select '1 year 2 mons 3 days

Re: [HACKERS] 8.3 vs HEAD difference in Interval output?

2008-10-09 Thread Kenneth Marshall
On Thu, Oct 09, 2008 at 02:47:24PM -0500, Kevin Grittner wrote: Kenneth Marshall [EMAIL PROTECTED] wrote: Even more surprising is the behavior for interval(1) here: [ some context with nonsurprising examples removed ...] ccdev=# select '1 year 2 mons 3 days 04:05:06.64321

Re: [HACKERS] CLUSTER, REINDEX, VACUUM in read only transaction?

2008-10-10 Thread Kenneth Marshall
On Fri, Oct 10, 2008 at 09:41:39AM -0400, Tom Lane wrote: So I was looking for other omissions in utility.c, and I noticed that check_xact_readonly() doesn't reject CLUSTER, REINDEX, or VACUUM. Now the notion of read only that we're trying to enforce is pretty weak (I think it's effectively no

Re: [HACKERS] minimal update

2008-10-22 Thread Kenneth Marshall
On Wed, Oct 22, 2008 at 06:05:26PM -0400, Tom Lane wrote: Simon Riggs [EMAIL PROTECTED] writes: On Wed, Oct 22, 2008 at 3:24 PM, Tom Lane [EMAIL PROTECTED] wrote: Minimal really fails to convey the point here IMHO. How about something like suppress_no_op_updates_trigger? I think it

Re: [HACKERS] WIP patch: convert SQL-language functions to return tuplestores

2008-10-28 Thread Kenneth Marshall
On Tue, Oct 28, 2008 at 09:28:38AM -0400, Tom Lane wrote: Simon Riggs [EMAIL PROTECTED] writes: On Sun, 2008-10-26 at 21:49 -0400, Tom Lane wrote: So I'm concluding that we can easily afford to switch to tuplestore-always operation, especially if we are willing to put any effort into

Re: [HACKERS] [PATCHES] updated hash functions for postgresql v1

2008-11-04 Thread Kenneth Marshall
lookup3 is 5n+20 where CRC is 9n+3. Regards, Ken On Tue, 4 Nov 2008, Kenneth Marshall wrote: Sorry about the delay for this update to the new hash index implementation. I was trying to get the WAL logging in place and forgot to post the actual patch. The WAL for hash indexes will need to wait

Re: [HACKERS] [PATCHES] updated hash functions for postgresql v1

2008-11-04 Thread Kenneth Marshall
Oleg, Here is a little more information on the use of CRC32 as a hash function, with some warning caveats: http://home.comcast.net/~bretm/hash/8.html Regards, Ken On Tue, Nov 04, 2008 at 03:15:44PM -0600, Kenneth Marshall wrote: On Tue, Nov 04, 2008 at 11:32:47PM +0300, Oleg Bartunov wrote

Re: [HACKERS] [PATCHES] updated hash functions for postgresql v1

2008-11-04 Thread Kenneth Marshall
Sorry about the delay for this update to the new hash index implementation. I was trying to get the WAL logging in place and forgot to post the actual patch. The WAL for hash indexes will need to wait for 8.5, but I did want to add back in the piece of the Bob Jenkins 2006 hash function that was

[HACKERS] Tests citext casts

2008-11-05 Thread Kenneth Marshall
I installed and ran the citext tests both with and without the patch and had failures both times. The patch applied cleanly and the make;make install completed without errors. I have attached the two regression.diffs files, one without the patch applied and the other with the patch. Regards, Ken

Re: [RRR] [HACKERS] Tests citext casts

2008-11-05 Thread Kenneth Marshall
On Wed, Nov 05, 2008 at 09:04:04AM -0800, David E. Wheeler wrote: On Nov 5, 2008, at 6:40 AM, Kenneth Marshall wrote: I installed and ran the citext tests both with and without the patch and had failures both times. The patch applied cleanly and the make;make install completed without errors

Re: [RRR] [HACKERS] Tests citext casts

2008-11-07 Thread Kenneth Marshall
On Fri, Nov 07, 2008 at 10:15:17AM -0800, David E. Wheeler wrote: On Nov 5, 2008, at 12:34 PM, Kenneth Marshall wrote: I am using the anonymous CVS repository, it returns the following information in pg_catalog.pg_settings: What is lc_collate set to? % show lc_collate; FWIW, I just ran

Re: [HACKERS] [RRR] Tests citext casts - reviewed

2008-11-07 Thread Kenneth Marshall
The patch for the citext tests applied to module cleanly and the patched files resulted in a clean make installcheck run for the citext module. My previous problem was the result of not testing with a C locale database. This patch is ready to be applied. Regards, Ken Marshall -- Sent via

Re: [HACKERS] WIP parallel restore patch

2008-11-20 Thread Kenneth Marshall
Okay, I have had a chance to run some timing benchmarks. Here are my results for the parallel pg_restore patch: Ken -- Server settings: max_connections = 100 # (change requires restart) shared_buffers = 256MB

Re: [HACKERS] WIP parallel restore patch

2008-11-20 Thread Kenneth Marshall
On Thu, Nov 20, 2008 at 02:26:14PM -0500, Andrew Dunstan wrote: Kenneth Marshall wrote: Okay, I have had a chance to run some timing benchmarks. Here are my results for the parallel pg_restore patch: Ken -- Server settings

Re: [HACKERS] cvs head initdb hangs on unixware

2008-12-09 Thread Kenneth Marshall
Would it be reasonable to turn of optimization for this file? Ken On Tue, Dec 09, 2008 at 05:47:47PM +0100, [EMAIL PROTECTED] wrote: On Tue, 9 Dec 2008, Tom Lane wrote: Date: Tue, 09 Dec 2008 09:23:06 -0500 From: Tom Lane [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: Heikki Linnakangas

Re: [HACKERS] Union test case broken in make check?

2009-11-19 Thread Kenneth Marshall
Without an order by, the order is not defined. The answers are the same but the test gives a false failure because of the lack of ordering. Regards, Ken On Thu, Nov 19, 2009 at 07:54:30PM -0500, Emmanuel Cecchet wrote: Tom Lane wrote: Andrew Dunstan and...@dunslane.net writes: Emmanuel

Re: [HACKERS] Avoiding bad prepared-statement plans.

2010-02-19 Thread Kenneth Marshall
On Thu, Feb 18, 2010 at 08:31:05PM -0600, David Christensen wrote: On Feb 18, 2010, at 2:19 PM, Pierre C wrote: What about catching the error in the application and INSERT'ing into the current preprepare.relation table? The aim would be to do that in dev or in pre-prod environments, then

Re: [HACKERS] pgbouncer + psql 9.0a4

2010-02-26 Thread Kenneth Marshall
Hi Garick, Add an ignore_startup_parameters to your pgbouncer.ini file with application_name. Cheers, Ken On Fri, Feb 26, 2010 at 11:26:23AM -0500, Garick Hamlin wrote: I was just trying out 9.0a4 and I noticed. That I can't connect to pgbouncer with psql from 9.0a4 as a result of the set

Re: [HACKERS] Can we still trust plperl?

2010-03-11 Thread Kenneth Marshall
On Thu, Mar 11, 2010 at 09:31:46AM -0500, Andrew Dunstan wrote: Last night my attention was drawn to this: http://search.cpan.org/~timb/PostgreSQL-PLPerl-Injector-1.002/lib/PostgreSQL/PLPerl/Injector.pm I'm wondering if we can reasonably continue to support plperl as a trusted language, or

[HACKERS] construct_array() use with PQexec with binary data

2010-05-05 Thread Kenneth Marshall
Dear PostgreSQL development community, I am working on adapting a regular PQexec() call to use binary transmission of the parameters. One of the parameters is an array of BIGINT. Looking in include/utils/array.h, it appears that construct_array() will do exactly what I need to get an array to

Re: [HACKERS] Generating Lots of PKs with nextval(): A Feature Proposal

2010-05-14 Thread Kenneth Marshall
Hi Peter, All you need to do is define your own sequence with an increment of 500. Look at: http://www.postgresql.org/docs/8.4/static/sql-createsequence.html Regards, Ken On Fri, May 14, 2010 at 02:56:18PM -0400, Peter Crabtree wrote: Recently, in preparation for migrating an application to

Re: [HACKERS] BYTEA / DBD::Pg change in 9.0 beta

2010-05-18 Thread Kenneth Marshall
On Tue, May 18, 2010 at 03:26:17PM -0600, Alex Hunsaker wrote: On Tue, May 18, 2010 at 15:20, Jesper Krogh jes...@krogh.cc wrote: On 2010-05-18 23:12, Alex Hunsaker wrote: set bytea_output 'escape'; That was it. Knowing what the problem was I had no problem finding it in the release

Re: [HACKERS] BYTEA / DBD::Pg change in 9.0 beta

2010-05-19 Thread Kenneth Marshall
On Wed, May 19, 2010 at 10:54:01AM -0400, Robert Haas wrote: On Wed, May 19, 2010 at 10:17 AM, Stefan Kaltenbrunner ste...@kaltenbrunner.cc wrote: On 05/19/2010 08:13 AM, Tom Lane wrote: Bernd Helmle maili...@oopsware.de writes: --On 18. Mai 2010 23:20:26 +0200 Jesper Krogh jes...@krogh.cc

Re: [HACKERS] Reduction in WAL for UPDATEs

2007-03-28 Thread Kenneth Marshall
On Wed, Mar 28, 2007 at 08:07:14AM +0100, Simon Riggs wrote: On Tue, 2007-03-27 at 20:48 -0400, Tom Lane wrote: Simon Riggs [EMAIL PROTECTED] writes: It seems possible to reduce overall WAL volume by roughly 25% on common workloads by optimising the way UPDATE statements generate WAL.

Re: [PATCHES] [HACKERS] Full page writes improvement, code update

2007-04-25 Thread Kenneth Marshall
On Wed, Apr 25, 2007 at 10:00:16AM +0200, Zeugswetter Andreas ADI SD wrote: 1) To deal with partial/inconsisitent write to the data file at crash recovery, we need full page writes at the first modification to pages after each checkpoint. It consumes much of WAL space. We need

Re: [HACKERS] todo: Hash index creation

2007-06-27 Thread Kenneth Marshall
On Wed, Jun 27, 2007 at 08:36:54PM -0400, Tom Lane wrote: Heikki Linnakangas [EMAIL PROTECTED] writes: [EMAIL PROTECTED] wrote: Is anyone currently working on this TODO item? During index creation, pre-sort the tuples to improve build speed If you want to work on hash indexes, though,

Re: [HACKERS] todo: Hash index creation

2007-07-07 Thread Kenneth Marshall
On Thu, Jul 05, 2007 at 12:26:45PM +0100, Heikki Linnakangas wrote: Kenneth Marshall wrote: I definitely agree with Tom's assessment. If we cannot need to make the hash index as performant as it is in theory, none of the other refinements are worth it. You would need to use BTree if you were

Re: [HACKERS] Query plan and execution time of a query

2007-07-21 Thread Kenneth Marshall
Mickael, Not knowing your query, it sounds like your method is working correctly. It is quite normal to have the initial query take longer than subsequent queries. This is a cache effect and is what databases, in general, strive for performance-wise. I suspect that the second time you run the

Re: [HACKERS] tsearch2 in PostgreSQL 8.3?

2007-08-14 Thread Kenneth Marshall
On Tue, Aug 14, 2007 at 03:15:44PM -0400, Alvaro Herrera wrote: Bruce Momjian wrote: A lot of work has been done to try to get /contrib/tsearch2 into the core backend for 8.3, but we have hit a roadblock in how to handle multiple text search configurations. (FYI, the documentation is at

[HACKERS] Hash index todo list item

2007-09-02 Thread Kenneth Marshall
Dear PostgreSQL Hackers: After following the hackers mailing list for quite a while, I am going to start investigating what will need to be done to improve hash index performance. Below are the pieces of this project that I am currently considering: 1. Characterize the current hash index

Re: [HACKERS] Hash index todo list item

2007-09-03 Thread Kenneth Marshall
On Sun, Sep 02, 2007 at 10:41:22PM -0400, Tom Lane wrote: Kenneth Marshall [EMAIL PROTECTED] writes: ... This is the rough plan. Does anyone see anything critical that is missing at this point? Sounds pretty good. Let me brain-dump one item on you: one thing that hash currently has over

Re: [HACKERS] Hash index todo list item

2007-09-03 Thread Kenneth Marshall
On Mon, Sep 03, 2007 at 10:33:54AM +0100, Simon Riggs wrote: This is the rough plan. Does anyone see anything critical that is missing at this point? Please send me any suggestions for test data and various performance test ideas, since I will be working on that first. Sounds good.

Re: [HACKERS] Hash index todo list item

2007-09-03 Thread Kenneth Marshall
On Mon, Sep 03, 2007 at 05:20:34PM -0700, Ben Tilly wrote: That raises a very random thought. One of the nicer features of Oracle is the ability to have function-based indexes. So you could index, say, trim(lower(person.name)). There are a *lot* of practical situations where that comes in

Re: [HACKERS] Hash index todo list item

2007-09-05 Thread Kenneth Marshall
On Sun, Sep 02, 2007 at 01:04:04PM -0500, Kenneth Marshall wrote: Dear PostgreSQL Hackers: After following the hackers mailing list for quite a while, I am going to start investigating what will need to be done to improve hash index performance. Below are the pieces of this project that I

Re: [HACKERS] Hash index todo list item

2007-09-06 Thread Kenneth Marshall
On Thu, Sep 06, 2007 at 11:53:45AM -0400, Mark Mielke wrote: Hannu Krosing wrote: One approahc is not to mix hashes, but to partition the hash, so that each column gets its N bits in the hash. How does that help? You still need all the keys to find out which bucket to look in.

Re: [HACKERS] Hash index todo list item

2007-09-07 Thread Kenneth Marshall
On Fri, Sep 07, 2007 at 09:50:07AM -0400, Mark Mielke wrote: Kenneth Marshall wrote: On Thu, Sep 06, 2007 at 11:56:25PM -0700, Neil Conway wrote: You might find this patch useful: http://archives.postgresql.org/pgsql-patches/2005-05/msg00164.php ... Unfortunately, the patch doesn't

Re: [HACKERS] Hash index todo list item

2007-09-07 Thread Kenneth Marshall
On Fri, Sep 07, 2007 at 12:55:37PM +0100, Heikki Linnakangas wrote: Neil Conway wrote: You might find this patch useful: http://archives.postgresql.org/pgsql-patches/2005-05/msg00164.php Oh, I had forgot about that. It implements the just store the hash in the index idea; it

Re: [HACKERS] Hash index todo list item

2007-09-07 Thread Kenneth Marshall
On Thu, Sep 06, 2007 at 11:56:25PM -0700, Neil Conway wrote: On Sun, 2007-02-09 at 13:04 -0500, Kenneth Marshall wrote: 2. Evaluate the performance of different hash index implementations and/or changes to the current implementation. My current plan is to keep the implementation

Re: [HACKERS] Hash index todo list item

2007-09-07 Thread Kenneth Marshall
On Thu, Sep 06, 2007 at 11:56:25PM -0700, Neil Conway wrote: On Sun, 2007-02-09 at 13:04 -0500, Kenneth Marshall wrote: 2. Evaluate the performance of different hash index implementations and/or changes to the current implementation. My current plan is to keep the implementation

Re: [HACKERS] Hash index todo list item

2007-09-07 Thread Kenneth Marshall
On Fri, Sep 07, 2007 at 10:36:41AM -0400, Brian Hurt wrote: Kenneth Marshall wrote: I understand that a hash value is a many-to-one mapping. That is the point of the flag in the index. The flag means that there is only one item in the heap corresponding to that hash value. In this case we

Re: [HACKERS] Hash index todo list item

2007-09-07 Thread Kenneth Marshall
On Fri, Sep 07, 2007 at 11:08:13AM -0400, Brian Hurt wrote: Kenneth Marshall wrote: How likely is it that you will get a hash collision, two strings that are different that will hash to the same value? To avoid this requires a very large hash key (128 bits, minimum)- otherwise you

Re: [HACKERS] Hash index todo list item

2007-09-07 Thread Kenneth Marshall
On Fri, Sep 07, 2007 at 10:30:30AM -0400, Mark Mielke wrote: Kenneth Marshall wrote: I understand that a hash value is a many-to-one mapping. That is the point of the flag in the index. The flag means that there is only one item in the heap corresponding to that hash value. In this case we

Re: [HACKERS] Hash index todo list item

2007-09-08 Thread Kenneth Marshall
On Fri, Sep 07, 2007 at 10:36:41AM -0400, Brian Hurt wrote: Kenneth Marshall wrote: I understand that a hash value is a many-to-one mapping. That is the point of the flag in the index. The flag means that there is only one item in the heap corresponding to that hash value. In this case we

Re: [HACKERS] Hash index todo list item

2007-09-08 Thread Kenneth Marshall
On Sat, Sep 08, 2007 at 05:14:09PM -0400, Mark Mielke wrote: Kenneth Marshall wrote: Continuing this train of thought While it would make sense for larger keys to store the hash in the index, if the key is smaller, particularly if it is of fixed size, it would make sense to store the key

Re: [HACKERS] Hash index todo list item

2007-09-09 Thread Kenneth Marshall
On Sun, Sep 02, 2007 at 10:41:22PM -0400, Tom Lane wrote: Kenneth Marshall [EMAIL PROTECTED] writes: ... This is the rough plan. Does anyone see anything critical that is missing at this point? Sounds pretty good. Let me brain-dump one item on you: one thing that hash currently has over

Re: [HACKERS] Hash index todo list item

2007-09-25 Thread Kenneth Marshall
On Thu, Sep 20, 2007 at 05:12:45PM -0700, Tom Raney wrote: We are pleased to announce an upcoming patch to the hash index code which improves build time and index size, based on this item in the TODO list: During index creation, pre-sort the tuples to improve build speed

Re: [HACKERS] Hash index todo list item

2007-09-25 Thread Kenneth Marshall
On Tue, Sep 25, 2007 at 03:35:47PM +0100, Gregory Stark wrote: Kenneth Marshall [EMAIL PROTECTED] writes: On Thu, Sep 20, 2007 at 05:12:45PM -0700, Tom Raney wrote: Using our implementation, build times and index sizes are comparable with btree index build times and index sizes

Re: [HACKERS] Fixing geometic calculation

2009-08-07 Thread Kenneth Marshall
On Fri, Aug 07, 2009 at 11:29:47PM +1000, Paul Matthews wrote: Let us consider the ordering of real numbers in postgres. As you can see from the results below it has clearly returned the correct results. select( 1. = 1.0002 ); = f select( 1.

Re: [HACKERS] Fixing geometic calculation

2009-08-07 Thread Kenneth Marshall
On Fri, Aug 07, 2009 at 09:12:34AM -0500, Kenneth Marshall wrote: On Fri, Aug 07, 2009 at 11:29:47PM +1000, Paul Matthews wrote: Let us consider the ordering of real numbers in postgres. As you can see from the results below it has clearly returned the correct results. select

  1   2   3   >