Hi,All.
The problem has occurred in my customer.
poor performance with Context Switch Storm occurred
with the following composition.
Usually, CS is about 5000, WIPS=360.
when CSStorm occurrence, CS is about 10, WIPS=60 or less.
(WIPS = number of web interactions per second)
It is under
I'd like to submit PL/Java into core for 8.2 if possible. Personally, I see the following
action items to make it happen:
1. A hackers discussion to resolve any issues with the submission.
Provided that #1 has a positive outcome:
2. The PL/Java CVS must be moved from gborg and become part of
Katsuhiko Okano [EMAIL PROTECTED] wrote
The problem has occurred in my customer.
poor performance with Context Switch Storm occurred
with the following composition.
Usually, CS is about 5000, WIPS=360.
when CSStorm occurrence, CS is about 10, WIPS=60 or less.
Intel Xeon 3.0GHz*4(2CPU
I am in the process of cleaning up the include files. Each include file
should compile on its own, and we should only include files we need.
Also, PostgreSQL include files should be in alphabetical order.
I do this cleanup every few years.
--
Bruce Momjian [EMAIL PROTECTED]
EnterpriseDB
Thomas Hallgren [EMAIL PROTECTED] writes:
I'd like to submit PL/Java into core for 8.2 if possible. Personally, I see
the following
action items to make it happen:
What about licensing issues? Does PL/Java work with any entirely-open-source
JVMs? If not, what is the legal situation for
Just having a standby mode that survived shutdown/startup would be a nice
start...
I also do the blocking-restore-command technique, which although workable,
has a bit of a house-of-cards feel to it sometimes.
On 7/10/06 5:40 PM, Florian G. Pflug [EMAIL PROTECTED] wrote:
Merlin Moncure
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
Katsuhiko Okano wrote:
I suspected conflict of BufMappingLock.
but, collected results are seen,
occurrence of CSStorm and the increase of BufMappingLock counts
seem not to correspond.
Instead, SubtransControlLock and SubTrans were increasing.
I do not understand what in the cause of
Currently, we restrict btree index tuples to a size that ensures three of
them will fit on a page. The motivation for this is the following two
considerations:
1. In a non-rightmost page, we need to include a high key, or page
boundary key, that isn't one of the useful data keys.
2. In a
Marc G. Fournier wrote:
If this is chosen as the preferred path, we could get the list bot to
add Reply-To: pghackers in pgsql-patches postings to help push
discussions there. I'd vote for doing the same in pgsql-committers,
which also gets its share of non-null discussion content.
that is a
Tom,
Tom Lane wrote:
IOW pljava is (already) bigger than the other four PLs put together.
I'm inclined to think that pljava is best off staying as a separate
project.
I was very confused some recent PL/Java versions can't be compiled
because of PostgreSQL internal changes.
If people think
Katsuhiko,
Have you tried turning HT off? HT is not generally considered (even by
Intel) a good idea for database appplications.
--Josh
---(end of broadcast)---
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose
Thomas Hallgren wrote:
5. I'll need committer rights to the PL/Java part in order to maintain
it.
Does our CVS setup cater for seggregated rights like this? Or would that
be done on a trust basis?
cheers
andrew
---(end of broadcast)---
On Tue, Jul 11, 2006 at 01:50:40AM +0300, Tzahi Fadida wrote:
As i understand rowids, i.e ctids, are supposed to allow for fast access to
the tables. I don't see the rational, for example, when casting some
attributes, to blank the ctid. So it is not exactly the same, but it still
came from
Tom,
What about licensing issues? Does PL/Java work with any entirely-open-source
JVMs? If not, what is the legal situation for distributing PG+PL/Java?
Actually, Sun has re-licensed the JRE to make it OSS-compatible (it's
now available for Debian, for example) They're doing a Java
Tom,
Obviously a tree containing many such pages would be awfully inefficient
to search, but I think a more common case is that there are a few wide
entries in an index of mostly short entries, and so pushing the hard
limit up a little would add some flexibility with little performance
cost in
On Tue, Jul 11, 2006 at 11:21:54PM +0900, Satoshi Nagayasu wrote:
Tom,
Tom Lane wrote:
IOW pljava is (already) bigger than the other four PLs put
together.
I'm inclined to think that pljava is best off staying as a
separate project.
I was very confused some recent PL/Java versions
Andrew Dunstan wrote:
Marc G. Fournier wrote:
If this is chosen as the preferred path, we could get the list bot to
add Reply-To: pghackers in pgsql-patches postings to help push
discussions there. I'd vote for doing the same in pgsql-committers,
which also gets its share of non-null
Andrew Dunstan wrote:
Thomas Hallgren wrote:
5. I'll need committer rights to the PL/Java part in order to maintain
it.
Does our CVS setup cater for seggregated rights like this? Or would that
be done on a trust basis?
Trust.
--
Bruce Momjian [EMAIL PROTECTED]
EnterpriseDB
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Andrew Dunstan
Sent: 11 July 2006 15:27
To: Thomas Hallgren
Cc: pgsql-hackers@postgresql.org
Subject: Re: [HACKERS] Three weeks left until feature freeze
Thomas Hallgren wrote:
5. I'll
Andrew Dunstan wrote:
Bruce Momjian wrote:
Don't be surprised if there are objections - this is one of those things
like emacs vs vi that stirs up religious debate.
If we change Reply-To:, does it prevent replies to the original author?
If so, that seems like a problem,
Bruce Momjian wrote:
Andrew Dunstan wrote:
Thomas Hallgren wrote:
5. I'll need committer rights to the PL/Java part in order to maintain
it.
Does our CVS setup cater for seggregated rights like this? Or would that
be done on a trust basis?
Trust.
And pgsql-committers
On Tuesday 11 July 2006 17:27, Martijn van Oosterhout wrote:
On Tue, Jul 11, 2006 at 01:50:40AM +0300, Tzahi Fadida wrote:
As i understand rowids, i.e ctids, are supposed to allow for fast access
to the tables. I don't see the rational, for example, when casting some
attributes, to blank
Does our CVS setup cater for seggregated rights like this? Or
would that
be done on a trust basis?
No, I don't believe you can do this with CVS at all. We'd need something
like SVN/WebDAV to be able to grant write access just to specific parts
of the tree to different people.
It is
Josh Berkus wrote:
Tom,
What about licensing issues? Does PL/Java work with any
entirely-open-source
JVMs? If not, what is the legal situation for distributing PG+PL/Java?
Actually, Sun has re-licensed the JRE to make it OSS-compatible (it's
now available for Debian, for example)
Bruce Momjian wrote:
Don't be surprised if there are objections - this is one of those things
like emacs vs vi that stirs up religious debate.
If we change Reply-To:, does it prevent replies to the original author?
If so, that seems like a problem, particularly if they are not
Josh Berkus wrote:
I'm inclined to think that pljava is best off staying as a separate
project.
I disagree. One of the things I'm asked by every single tech market
analyst, after replication clustering, is whether we have support for
procedural Java. So it's something large-scale
On Monday 10 July 2006 17:06, Josh Berkus wrote:
All,
At the request of Dave Page, here's the semi-final list after looking at
the code:
To be killed:
adddepends
tips
mSQL-interface
To be honest I don't know why people are against throwing the code on
pgfoundry with a
Andrew Dunstan [EMAIL PROTECTED] writes:
Marc G. Fournier wrote:
If nobody objects within, say, the next 24 hours ... ? I'll enabled that
one
both ...
Don't be surprised if there are objections - this is one of those things like
emacs vs vi that stirs up religious debate.
Indeed.
On Mon, 10 Jul 2006, Josh Berkus wrote:
To be migrated to pgFoundry:
dbmirror (need owner)
I'll volunteer for this if no one else steps forward. I'm not planning on
making any significant chances to dbmirror at this point stage but I can
look after for the pgfoundry project.
* Greg Stark ([EMAIL PROTECTED]) wrote:
I have the additional complaint that this doesn't actually solve most of my
original complaints and might reduce the pressure to find a better solution.
The patches announcements themselves would still be basically invisible within
the community.
I'm
* Josh Berkus (josh@agliodbs.com) wrote:
Actually, Sun has re-licensed the JRE to make it OSS-compatible (it's
now available for Debian, for example) They're doing a Java licensing
session at OSCON if you have any specific questions, or I can ping the
Java Licensing Guru directly. But
Robert,
To be honest I don't know why people are against throwing the code on
pgfoundry with a hefty readme saying that the code is unmaintained and what
it's build status is on various versions
... because we don't want to litter pgFoundry with dead, broken projects
which nobody uses and
Robert,
Given the current number of projects that have no code / files / anything
associated with them on pgfoundry/gborg right now, this argument rings a
little hollow.
If you're so keen to add to the problem, you can have my spot as
pgfoundry admin.
Otherwise, the rule that the
David,
It's good to integrate things with the core as needed. What plans do
we have to integrate PL/J?
None, if the PL/J team doesn't speak up. So far I have yet to see a
request for PL/J or even a release notice.
--Josh
---(end of
On Tuesday 11 July 2006 12:55, Josh Berkus wrote:
Robert,
To be honest I don't know why people are against throwing the code on
pgfoundry with a hefty readme saying that the code is unmaintained and
what it's build status is on various versions
... because we don't want to litter
Snowman,
Uhh.. Let's not go overboard here on exactly what Debian has done with
Sun's JVM. Technically, Sun's JVM is *not* part of Debian. The license
is (and even this is hotly debated...) acceptable enough for Debian's
ftp-masters to allow the Sun JVM to be distributed off Debian servers
On 7/10/06, Josh Berkus josh@agliodbs.com wrote:
All,
userlock (Merlin)
Ok, I will update the project description and maintain it. userlock
is a great feature, and I tried contacting the original author to get
him to relicense the project but could never get a hold of him. To be
Andrew Dunstan [EMAIL PROTECTED] writes:
Thomas Hallgren wrote:
5. I'll need committer rights to the PL/Java part in order to maintain
it.
Does our CVS setup cater for seggregated rights like this? Or would that
be done on a trust basis?
No, and yes. However, I don't have a problem with
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
NotDashEscaped: You need GnuPG to verify this message
One thing that came up in the discussion here was the idea of a
weekly (or other time period) digest of patches posts, stripped
of attachments, but with a link to the patches email, which will
Bruce Momjian [EMAIL PROTECTED] writes:
Let's add the author and the hackers list to the reply-to.
I think reply-to is just a single address. It may work in some mailers though.
Regardless the issue is that someone may send a personal message and be
surprised when it's broadcast. You can
Stephen,
You seemed to use the recent change in status of Sun's JVM (at least in
part with regard to Debian...) as justification of your statement that
it's OSS-compatible..
Are you going to be at OSCON? Sun's hosting a BOF to discuss exactly
this issue.
--Josh
On Tuesday 11 July 2006 14:05, Josh Berkus wrote:
Robert,
Given the current number of projects that have no code / files / anything
associated with them on pgfoundry/gborg right now, this argument rings a
little hollow.
If you're so keen to add to the problem, you can have my spot as
Currently, functional indexes can be used by a query that explicitly
mentions the exact phrasing of the functional index within the WHERE
clause.
IMHO it is feasible to extend the range of WHERE clauses for which
functional indexes can be used by using implication, much in the same
way that we
Robert,
No need to fly off the handle there Josh.
I was hoping that you'd take me up on it in a rash moment.
No code, or no active code development?
No code was the rule we discussed. Other stuff would be a matter for
discussion. The idea was that pgfoundry was supposed to be
On Mon, 2006-07-10 at 19:34 +0200, Florian G. Pflug wrote:
This methods seems to work, but it is neither particularly fool-proof nor
administrator friendly. It's not possible e.g. to reboot the slave without
postgres
abortint the recovery, and therefor processing all wals generated since the
On Tuesday 11 July 2006 13:49, Andrew Dunstan wrote:
Joshua D. Drake wrote:
... and before you say it, No. I do not wear a tie.
Maybe you need to ... ;-)
/me bows before the gods who thoust commit.
cheers
andrew
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support:
Joshua D. Drake wrote:
... and before you say it, No. I do not wear a tie.
Maybe you need to ... ;-)
cheers
andrew
---(end of broadcast)---
TIP 3: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faq
Hi :)
powered by the great talk for newbie hackers by Neil Conway Gavin
Sherry [1] at the anniversary summit I sneaked into the Code Sprint
and started working a bit on a Todo from Neil's Code Sprint Page:
Add a psql option to have it submit SELECT queries via a cursor,
to avoid excessive
* Josh Berkus (josh@agliodbs.com) wrote:
I think I can tell which side of the debate you were on.
The debate was regarding Sun's JVM being distributed by Debian at all...
There wasn't any debate regarding it's free vs. non-free status so far
as I'm aware. I don't believe there was ever any
On Tue, Jul 11, 2006 at 03:53:26PM -0400, Josh Berkus wrote:
All I am saying is that it couldn't hurt to put the information out
there... we're not hurting for disk space and none of this stuff appears
inherently wrong, just outdated, but it might still prove useful for some
people.
Hi Hannu,
Hannu Krosing wrote:
Maybe this functionality could be lifted out of PL/Java and made
available to all PL-s ?
At least at some API level.
I think that what could be shared are the ideas and the semantics. The
API's that the backend currently expose will give you what's needed to
On Tuesday 11 July 2006 09:59, Tom Lane wrote:
Andrew Dunstan [EMAIL PROTECTED] writes:
Thomas Hallgren wrote:
5. I'll need committer rights to the PL/Java part in order to maintain
it.
Does our CVS setup cater for seggregated rights like this? Or would that
be done on a trust basis?
On Tue, 11 Jul 2006, Bruce Momjian wrote:
Andrew Dunstan wrote:
Marc G. Fournier wrote:
If this is chosen as the preferred path, we could get the list bot to
add Reply-To: pghackers in pgsql-patches postings to help push
discussions there. I'd vote for doing the same in pgsql-committers,
Robert,
I really don't see how this will actually cause you any extra effort, but if
you want to plug my name on there after you move it, that's fine with me.
I meant maintain it, not just leave it there to age like a bad cheese.
If it's going to be dead code, it can do so in the FTP
On Tue, 11 Jul 2006, Josh Berkus wrote:
I disagree. One of the things I'm asked by every single tech market
analyst, after replication clustering, is whether we have support for
procedural Java. So it's something large-scale users want. If PL/Tcl
belongs in the back end, then so does
Simon Riggs [EMAIL PROTECTED] writes:
...
- add a new boolean to pg_operator to allow us to define which operators
offer true equality
...
This would be useful for other purposes too, as we keep coming up
against what's the equality operator for this datatype problems.
However, the
On Tuesday 11 July 2006 16:33, Josh Berkus wrote:
Robert,
I really don't see how this will actually cause you any extra effort, but
if you want to plug my name on there after you move it, that's fine with
me.
I meant maintain it, not just leave it there to age like a bad cheese.
If
Ühel kenal päeval, T, 2006-07-11 kell 08:38, kirjutas Andrew Rawnsley:
Just having a standby mode that survived shutdown/startup would be a nice
start...
I think that Simon Riggs did some work on this at the code sprint
yesterday.
I also do the blocking-restore-command technique, which
On Tue, Jul 11, 2006 at 06:28:31PM -, Greg Sabino Mullane wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
NotDashEscaped: You need GnuPG to verify this message
One thing that came up in the discussion here was the idea of a
weekly (or other time period) digest of patches posts,
Try again, today's patch solves the problem.
Stefan Kaltenbrunner wrote:
on IRC somebody mentioned that it took 34h to greate a GIN index (on a
tsvector) on a ~3 Million column table (wikipedia dump) with a
reasonable speced box (AMD 3400+).
After getting hold of a dump of said table (around
Ühel kenal päeval, T, 2006-07-11 kell 10:46, kirjutas Josh Berkus:
Tom,
Obviously a tree containing many such pages would be awfully inefficient
to search, but I think a more common case is that there are a few wide
entries in an index of mostly short entries, and so pushing the hard
Ühel kenal päeval, T, 2006-07-11 kell 17:34, kirjutas Thomas Hallgren:
Josh Berkus wrote:
Tom,
IOW pljava is (already) bigger than the other four PLs put together.
That is odd. Thomas?
It's not that odd really:
1. the mapping is strongly typed, i.e. each scalar type in
Ühel kenal päeval, T, 2006-07-11 kell 14:05, kirjutas Josh Berkus:
Robert,
Given the current number of projects that have no code / files / anything
associated with them on pgfoundry/gborg right now, this argument rings a
little hollow.
If you're so keen to add to the problem, you
Given the current number of projects that have no code / files / anything
associated with them on pgfoundry/gborg right now, this argument rings a
little hollow.
I would say that:
Given the current number of projects that have no code / files / anything
associated with them on
On Tue, 11 Jul 2006, Greg Stark wrote:
I have the additional complaint that this doesn't actually solve most of
my original complaints and might reduce the pressure to find a better
solution. The patches announcements themselves would still be basically
invisible within the community.
How
On Tuesday 11 July 2006 15:53, Josh Berkus wrote:
I'll grant that tips
doesn't look like much more than an article stub... it should probably be
moved to the new techdocs rather than pgfoundry.
That was what I started to do. Unfortunately, the README is
instrucitons for some SQL and
On 7/9/2006 8:32 AM, Martijn van Oosterhout wrote:
On Sat, Jul 08, 2006 at 05:47:33PM -0400, Jim Nasby wrote:
On Jul 6, 2006, at 11:02 AM, Phil Frost wrote:
I hope the above example is strong enough to elicit a comment from a
qualified developer. If it is not, consider that stored procedures
hello.
Do you have bgwriter on and what's the parameters? I read a theory somewhere
that bgwriter scan a large portion of memory and cause L1/L2 thrushing, so
with HT on, the other backends sharing the physical processor with it also
get thrashed ... So try to turn bgwriter off or turn HT
I've already added adddepends to pgFoundry (as Old PG Upgrade), since
people spoke up for it. I will assign one of them as admin of the
project (not sure who yet).
How is addepends in any way old pg upgrade??
---(end of broadcast)---
TIP 5:
No, I don't believe you can do this with CVS at all. We'd need something
like SVN/WebDAV to be able to grant write access just to specific parts
of the tree to different people.
You just use an on-commit script like cvsacl.
---(end of
Again, it's the same question. If *you* want to be the maintainer, I'll
put it on pgfoundry. Otherwise, you're asking me to be responsible for
the code because you don't want to throw it away.
Josh,
How about a general call for maintainers? Post it to general, hackers and
advocacy (maybe
I am trying to create an initialisation function that is called using
the preload_libraries option.
The purpose of this is to set up shared memory for Veil, independant of
postgres' own shared memory. Simple init functions work fine, but as
soon as I place calls to ShemAlloc, or LWLockAssign,
I'm afraid there's not much in the PL/Java type system that could be
generalized and shared. Perhaps if we had other languages with very
similar capabilities (like C# for instance) but even then I have some
doubts. The good news in my opinion is that if PL/Java would make it to
the core it
Tom Lane wrote:
Greg Stark [EMAIL PROTECTED] writes:
The SQL spec has some detailed discussion of some strange null behaviours.
BTW, Teodor Sigaev pointed out today that we are also doing array
comparisons (array_eq, array_cmp) wrong.
Seems to me like at least array_eq is correct (from
Charles Duffy [EMAIL PROTECTED] writes:
Their work_mem setting was rather large (100). We determined that when it
received SIGINT, the backend was always inside qsort(), so it wouldn't
call ProcessInterrupts() again until it finished this large in-memory
sort. Upon entering
Teodor Sigaev [EMAIL PROTECTED] wrote:
pluggable parameters for index.
I think, we may
can add to pg_opclass's definition method/parameter name and create some API
(may be, index specific) to propagate parameter's to module's interface
functions to index.
How abount adding a new option
On Tue, 2006-07-11 at 21:19 +0200, Chris Mair wrote:
One of the problems with this was that a user would expect psql to
work as usual (including all format and output option stuff) and
to do this properly most of the psql output code would need to be
refactored.
Even if the refactoring were
Marc G. Fournier wrote:
On Tue, 11 Jul 2006, Josh Berkus wrote:
I disagree. One of the things I'm asked by every single tech market
analyst, after replication clustering, is whether we have support for
procedural Java. So it's something large-scale users want. If PL/Tcl
belongs in
79 matches
Mail list logo