Re: [HACKERS] Patent warning about the Greenplum source code

On Fri, Oct 30, 2015 at 09:56:48AM +0100, Andres Freund wrote: > Hi, > > I don't really want to discuss patent issues publically. While we don't want to discuss patented ideas, the patent terms are an imporant topic here. > On 2015-10-30 04:47:35 -0400, Bruce Momjian wrote: > > However, while

Re: [HACKERS] pg_dump

> More specifically, I am not quite understanding the distinction > between "all cluster" and "global objects." all cluster is roles, tablespaces, databases with it's content. global objects is roles, tablespaces. > What do you have in mind on the implementation side? Do you think > pg_dump is

[HACKERS] Patent warning about the Greenplum source code

Some of you might have seen that the Greenplum database source code has been published: https://adtmag.com/articles/2015/10/28/greenplum-open-sourced.aspx under the Apache 2.0 license: http://www.apache.org/licenses/LICENSE-2.0 The source code has known patents owned by

Re: [HACKERS] Patent warning about the Greenplum source code

Hi, I don't really want to discuss patent issues publically. On 2015-10-30 04:47:35 -0400, Bruce Momjian wrote: > However, while the license defines and uses "Derivative Works", it does > not mention that in the patent grant clause. I assume this means that > patent grants do not apply to

Re: [HACKERS] Dangling Client Backend Process

On Tue, Oct 27, 2015 at 6:29 AM, Rajeev rastogi wrote: > On 23 October 2015 01:58, Robert Haas [mailto:robertmh...@gmail.com] Wrote: >>Well, I'm not buying this extra PostmasterIsAlive() call on every pass >>through the main loop. That seems more expensive than we can

Re: [HACKERS] Cross-check recent documentation changes

On 29 October 2015 at 15:35, Amit Langote wrote: > errmsg_plural() function determines whether to output the singular version > or the plural > ​Duh. Thanks Amit! Should have noticed the function-name change.​ -- Robins Tharakan

Re: [HACKERS] pg_dump LOCK TABLE ONLY question

Please take it as a very naive and basic approach :-) What could go wrong here? diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c index 36863df..57a50b5 100644 --- a/src/bin/pg_dump/pg_dump.c +++ b/src/bin/pg_dump/pg_dump.c @@ -5169,9 +5169,9 @@ getTables(Archive *fout,

Re: [HACKERS] September 2015 Commitfest

On Fri, Oct 30, 2015 at 2:02 PM, Robert Haas wrote: > On Fri, Oct 30, 2015 at 10:47 AM, Michael Paquier wrote: >> On Thu, Oct 22, 2015 at 9:21 AM, Andres Freund wrote: Among the five patches marked as ready for committer, one is a bug fix that should be back-patched (ahem). Shouldn't we

Re: [HACKERS] September 2015 Commitfest

On 2015-10-31 00:42:54 +0100, Michael Paquier wrote: > On Fri, Oct 30, 2015 at 2:02 PM, Robert Haas wrote: > > On Fri, Oct 30, 2015 at 10:47 AM, Michael Paquier wrote: > >> On Thu, Oct 22, 2015 at 9:21 AM, Andres Freund wrote: > Among the five patches marked as ready for committer, one is a

Re: [HACKERS] onlyvalue aggregate (was: First Aggregate Funtion?)

On Wed, Oct 28, 2015 at 5:50 PM, Marko Tiikkaja wrote: > Here's a patch for the aggregate function outlined by Corey Huinker in > CADkLM=foA_oC_Ri23F9PbfLnfwXFbC3Lt8bBzRu3=cb77g9...@mail.gmail.com . I > called it "onlyvalue", which is a horrible name, but I have nothing better > to

Re: [HACKERS] Getting sorted data from foreign server

If there is a collate clause in the ORDER BY, the server crashes with assertion +Assert(loc_cxt.state == FDW_COLLATE_NONE || +loc_cxt.state == FDW_COLLATE_SAFE); The assertion is fine as long as is_foreign_expr() tests only boolean expressions (appearing in quals). This patch

Re: [HACKERS] plpgsql - DECLARE - cannot to use %TYPE or %ROWTYPE for composite types

2015-10-19 9:52 GMT+02:00 Pavel Stehule : > Hi, > > We cannot to declare variable with referenced type on other composite > variable. This limit is probably artificial, because any composite type is > any type too in PostgreSQL. > > The issue: > > referencing on composite

[HACKERS] Re: [HACKERS] ExclusiveLock on PostgreSQL - Fabio Mendonça

On Wed, Oct 28, 2015 at 5:59 PM, Fabio Oliveira De Mendonca wrote: > I 've a process with 600.000 rows, for insert on table "A" with 130 columns > and I'm received the "Exclusivelock" error message, making lost some > rows during transaction. The insert of transaction

[HACKERS] ParallelContexts can get confused about which worker is which

While testing last night, I discovered a serious case of brain fade in parallel.c; the same conceptual mistake has also spread to nodeGather.c. parallel.c creates an array of ParallelWorkerInfo structures, which are defined like this: typedef struct ParallelWorkerInfo {

Re: [HACKERS] security_barrier view option type mistake in create view document

On Thu, Oct 29, 2015 at 6:02 AM, Haribabu Kommi wrote: > The security_barrier view option is classified as string in the create > view documentation. > But it is actually a boolean. The type is mentioned correctly in alter > view. Here I attached > the patch with the

Re: [HACKERS] Replication connection URI?

Heikki Linnakangas writes: > On 11/25/2014 05:11 PM, Heikki Linnakangas wrote: >> On 11/24/2014 06:05 PM, Alex Shulgin wrote: >>> Heikki Linnakangas writes: > > It appears that replication connection doesn't support URI but only the

Re: [HACKERS] September 2015 Commitfest

On Thu, Oct 22, 2015 at 9:21 AM, Andres Freund wrote: >> Among the five patches marked as ready for committer, one is a bug fix >> that should be back-patched (ahem). Shouldn't we move on with those >> entries first? > > I think at this point we essentially can just move all entries to the > next.

Re: [HACKERS] [ADMIN] Replication slots and isolation levels

> 30 окт. 2015 г., в 14:30, Robert Haas написал(а): > > On Thu, Oct 29, 2015 at 3:29 PM, Oleksii Kliukin wrote: >> Could it be a consequence of how REPEATABLE READ transactions handle >> snapshots? With REPEATABLE READ the snapshot is acquired only

Re: [HACKERS] [DESIGN] ParallelAppend

On Wed, Oct 28, 2015 at 3:55 PM, Kouhei Kaigai wrote: > At PGconf.EU, I could have a talk with Robert about this topic, > then it became clear we have same idea. > >> ++ >> |sub-plan | * Sub-Plan 1 ... Index Scan on p1 >> |index on *-> * Sub-Plan 2 ...

Re: [HACKERS] September 2015 Commitfest

Andres Freund writes: > On 2015-10-31 00:42:54 +0100, Michael Paquier wrote: >> On Fri, Oct 30, 2015 at 2:02 PM, Robert Haas wrote: >>> Gosh, that's a lot of stuff that didn't get reviewed. :-( >> Yep. > Yea, this is probably one of the worst commitfests ever from the point

Re: [HACKERS] Freeze avoidance of very large table.

On Thu, Oct 8, 2015 at 11:05 PM, Simon Riggs wrote: > > On 1 October 2015 at 23:30, Josh Berkus wrote: >> >> On 10/01/2015 07:43 AM, Robert Haas wrote: >> > On Thu, Oct 1, 2015 at 9:44 AM, Fujii Masao wrote: >> >> I wonder how

Re: [HACKERS] Parallel Seq Scan

On Wed, Oct 28, 2015 at 01:04:12AM +0100, Robert Haas wrote: > Well, OK. That's not strictly a correctness issue, but here's an > updated patch along the lines you suggested. > Finally, have setup_param_list set a new ParamListInfo field, > paramMask, to the parameters actually used in the

Re: [HACKERS] September 2015 Commitfest

On Sat, Oct 31, 2015 at 12:55 AM, Andres Freund wrote: > On 2015-10-31 00:42:54 +0100, Michael Paquier wrote: >> On Fri, Oct 30, 2015 at 2:02 PM, Robert Haas wrote: >> > On Fri, Oct 30, 2015 at 10:47 AM, Michael Paquier wrote: >> >> On Thu, Oct 22, 2015 at 9:21 AM, Andres

Re: [HACKERS] Request: pg_cancel_backend variant that handles 'idle in transaction' sessions

On Sat, Oct 31, 2015 at 2:50 AM, Merlin Moncure wrote: > > Idle hanging transactions from poorly written applications are the > bane of my existence. Several months back one of them took down one > of hour production websites for several hours. > > Unfortunately, the only way

Re: [HACKERS] [ADMIN] Replication slots and isolation levels

On 2015-10-30 13:42:19 +0100, Michael Paquier wrote: > On Fri, Oct 30, 2015 at 12:40 PM, Vladimir Borodin wrote: > > On Thu, Oct 29, 2015 at 3:29 PM, Oleksii Kliukin wrote: > >> Could it be a consequence of how REPEATABLE READ transactions handle > >> snapshots? With REPEATABLE READ the snapshot

Re: [HACKERS] [ADMIN] Replication slots and isolation levels

On Fri, Oct 30, 2015 at 12:40 PM, Vladimir Borodin wrote: > I still don’t fully understand why is it so (the problem occurs while > running only one SELECT-statement in READ COMMITED so only one snapshot is > taken), but if is expected behavior shouldn’t the documentation

Re: [HACKERS] WIP: Fix parallel workers connection bug in pg_dump (Bug #13727)

I'm still unclear on how to write regression tests for a connectivity bug. Are they necessary in this case? On Sun, Oct 25, 2015 at 5:55 PM, Zeus Kronion wrote: > Parallel workers were failing to connect to the database when running > pg_dump with a connection string. The

Re: [HACKERS] Dangling Client Backend Process

Robert Haas writes: > Hmm. ProcessInterrupts() signals some FATAL errors while the > connection is idle, and rumor has it that that works: the client > doesn't immediately read the FATAL error, but the next time it sends a > query, it tries to read from the connection and

Re: [HACKERS] Patch: Implement failover on libpq connect level.

On Fri, 30 Oct 2015 14:26:45 +0100 Robert Haas wrote: > On Thu, Oct 29, 2015 at 8:29 PM, Peter Eisentraut > wrote: > > That's true, but doesn't allowing every parameter to be multiply > specified greatly increase the implementation complexity for a

Re: [HACKERS] pg_dump

On 10/29/2015 03:51 PM, Tom Lane wrote: > We don't need hasty patches. What we need is a re-think of the division > of labor between pg_dump and pg_dumpall. Up to now, pg_dump has only been > charged with dumping/restoring the data "inside" an individual database, > not with handling any

Re: [HACKERS] September 2015 Commitfest

On Fri, Oct 30, 2015 at 10:47 AM, Michael Paquier wrote: > On Thu, Oct 22, 2015 at 9:21 AM, Andres Freund wrote: >>> Among the five patches marked as ready for committer, one is a bug fix >>> that should be back-patched (ahem). Shouldn't we move on with those >>>

Re: [HACKERS] Patch: Implement failover on libpq connect level.

On Thu, Oct 29, 2015 at 8:29 PM, Peter Eisentraut wrote: > On 10/28/15 4:18 AM, Victor Wagner wrote: >> On Mon, 26 Oct 2015 16:25:57 -0400 >> Peter Eisentraut wrote: >> >>> Also, this assumes that all the components other than host and port >>> are the same.

Re: [HACKERS] [ADMIN] Replication slots and isolation levels

> 30 окт. 2015 г., в 16:04, Robert Haas написал(а): > > On Fri, Oct 30, 2015 at 12:40 PM, Vladimir Borodin wrote: >> I still don’t fully understand why is it so (the problem occurs while >> running only one SELECT-statement in READ COMMITED so only one

Re: [HACKERS] Move PinBuffer and UnpinBuffer to atomics

On Thu, Oct 29, 2015 at 8:18 PM, Alexander Korotkov < a.korot...@postgrespro.ru> wrote: > On Thu, Sep 24, 2015 at 6:36 PM, Alexander Korotkov < > a.korot...@postgrespro.ru> wrote: > >> On Thu, Sep 24, 2015 at 6:32 PM, Andres Freund >> wrote: >> >>> On 2015-09-15 20:16:10

Re: [HACKERS] WIP: Fix parallel workers connection bug in pg_dump (Bug #13727)

On 30-10-2015 10:04, Zeus Kronion wrote: I'm still unclear on how to write regression tests for a connectivity bug. Are they necessary in this case? There aren't regression tests for pg_dump. However, your instructions are sufficient to demonstrate the bug. You could continue the thread in

Re: [HACKERS] Move PinBuffer and UnpinBuffer to atomics

Hi, Thanks for benchmarking! On 2015-10-30 16:28:22 +0300, Alexander Korotkov wrote: > pinunpin-cas-original-fix.patch is just original patch by Andres Freund > with fixed bug which causes hang. > Performance comparison on 72-cores Intel server in attached. On this > machine we see no regression

Re: [HACKERS] [ADMIN] Replication slots and isolation levels

On Fri, Oct 30, 2015 at 12:40 PM, Vladimir Borodin wrote: > On Thu, Oct 29, 2015 at 3:29 PM, Oleksii Kliukin wrote: >> Could it be a consequence of how REPEATABLE READ transactions handle >> snapshots? With REPEATABLE READ the snapshot is acquired only once at the >> beginning of a transaction; a

Re: [HACKERS] exposing pg_controldata and pg_config as functions

[2015082503-pgconfig_controldata.diff] I tried to build this, and the patch applies cleanly but then a ld error emerges: (The first four lines (about gram.y) are standard warnings; the error starts from the ld line) In file included from gram.y:14908:0: scan.c: In function

Re: [HACKERS] Did the "Full-text search in PostgreSQL in milliseconds" patches land?

On Thu, Oct 29, 2015 at 5:31 PM, Colin 't Hart wrote: > Hi, > > I've been reading > > wiki.postgresql.org/images/2/25/Full-text_search_in_PostgreSQL_in_milliseconds-extended-version.pdf > with interest and am wondering if these patches ever made it in to the > "official"

Re: [HACKERS] Dangling Client Backend Process

Andres Freund writes: > adding a parseInput(conn) into the loop yields the expected > FATAL: 57P01: terminating connection due to unexpected postmaster exit > Is there really any reason not to do that? Might work, but it probably needs some study: (a) is it safe (b) is this

Re: [HACKERS] Patch: Implement failover on libpq connect level.

On 30 October 2015 at 09:26, Robert Haas wrote: > > On Thu, Oct 29, 2015 at 8:29 PM, Peter Eisentraut wrote: > > On 10/28/15 4:18 AM, Victor Wagner wrote: > >> On Mon, 26 Oct 2015 16:25:57 -0400 > >> Peter Eisentraut wrote: > >> > >>>

Re: [HACKERS] extend pgbench expressions with functions

On Fri, Sep 18, 2015 at 10:21 AM, Fabien COELHO wrote: > > Hello Kyotaro-san, > > My description should have been obscure. Indeed the call tree is >> finite for *sane* expression node. But it makes infinit call for >> a value of expr->etype unknown by both evalDouble and >>

Re: [HACKERS] Dangling Client Backend Process

On 2015-10-30 09:48:33 -0400, Tom Lane wrote: > Robert Haas writes: > > Hmm. ProcessInterrupts() signals some FATAL errors while the > > connection is idle, and rumor has it that that works: the client > > doesn't immediately read the FATAL error, but the next time it

Re: [HACKERS] Dangling Client Backend Process

On 2015-10-30 10:57:45 -0400, Tom Lane wrote: > Andres Freund writes: > > adding a parseInput(conn) into the loop yields the expected > > FATAL: 57P01: terminating connection due to unexpected postmaster exit > > Is there really any reason not to do that? > > Might work, but

Re: [HACKERS] ExclusiveLock on PostgreSQL - Fabio Mendonça

Thanks Robert. I read the material link and did help me to take a new decision thank you. att. Fabio Mendonça De: Robert Haas Enviado: sexta-feira, 30 de outubro de 2015 07:49 Para: Fabio Oliveira De Mendonca Cc:

Re: [HACKERS] WIP: Fix parallel workers connection bug in pg_dump (Bug #13727)

On Fri, Oct 30, 2015 at 2:42 PM, Euler Taveira wrote: > On 30-10-2015 10:04, Zeus Kronion wrote: >> >> I'm still unclear on how to write regression tests for a connectivity >> bug. Are they necessary in this case? >> > There aren't regression tests for pg_dump. However, your

[HACKERS] Request: pg_cancel_backend variant that handles 'idle in transaction' sessions

Idle hanging transactions from poorly written applications are the bane of my existence. Several months back one of them took down one of hour production websites for several hours. Unfortunately, the only way to deal with them is to terminate the backend which is heavy handed and in some cases

Re: [HACKERS] [DESIGN] ParallelAppend

> On Wed, Oct 28, 2015 at 3:55 PM, Kouhei Kaigai wrote: > > At PGconf.EU, I could have a talk with Robert about this topic, > > then it became clear we have same idea. > > > >> ++ > >> |sub-plan | * Sub-Plan 1 ... Index Scan on p1 > >> |index on *-> *

Re: [HACKERS] Patch: Implement failover on libpq connect level.

On Fri, Oct 30, 2015 at 11:29:09AM -0400, Christopher Browne wrote: > On 30 October 2015 at 09:26, Robert Haas wrote: > > > > On Thu, Oct 29, 2015 at 8:29 PM, Peter Eisentraut wrote: > > > On 10/28/15 4:18 AM, Victor Wagner wrote: > > >> On Mon, 26 Oct

Re: [HACKERS] [PATCH v3] GSSAPI encryption support

Andreas, can you please weigh in here since your voice is important to this process? Robbie Harwood writes: > Andres Freund writes: > >> On 2015-10-22 16:47:09 +0900, Michael Paquier wrote: >>> Hm, and that's why you chose this way of going. My main

Re: [HACKERS] extend pgbench expressions with functions

Here is a v12 which implements the suggestions below. Should we not allow for functions taking 0 arguments? Since we're already into some math here, how about pi()? ;-) Hmmm, why not. I understand requiring at least 1 arg simplifies the code a bit, but right now it reports syntax error

Re: [HACKERS] Patch: Implement failover on libpq connect level.

On 10/30/2015 08:29 AM, Christopher Browne wrote: > I think it makes the feature WORSE. I am getting more and more convinced > that the Correct Solution is for this feature to be handled by submitting > multiple URIs, and my argument isn't even based on any aspects of > implementation complexity.

Re : Re: Re : Re: [HACKERS] UTF-32 support in PostgreSQL ?

Now I received the authorization to give you an answer to the WHY question! Because basicly, this project is classified TOP SECTRET. Well, we know then we have no real avantage to use UTF-32 in comparaison to UTF-8. But we need to establish a gateway between two huge networks. One network is

Re: [HACKERS] ALTER ... OWNER TO ... vs. ALTER DEFAULT PRIVILEGES

On Thu, Oct 29, 2015 at 10:31 PM, David Fetter wrote: > Had this been part of the original ALTER DEFAULT PRIVILEGES patch, > those privileges would simply have been applied. Since it wasn't, I'm > ass-u-me'ing that changing the default behavior to that is going to > cause

[HACKERS] Patch to install config/missing

Currently, config/missing isn't being installed. This can lead to confusing error messages, such as if Perl isn't found and something needs it [1]. Attached patch adds it to install and uninstall recipes. -- Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX Experts in Analytics, Data

Re: [HACKERS] [DOCS] max_worker_processes on the standby

On Thu, Oct 29, 2015 at 5:41 PM, Fujii Masao wrote: > I found another strange behavior on track_commit_timestamp. > Here are the steps to reproduce it. > > 1. Start the master and standby servers with track_commit_timestamp enabled. > Since committs is activated in

Re: [HACKERS] [ADMIN] Replication slots and isolation levels

On Thu, Oct 29, 2015 at 3:29 PM, Oleksii Kliukin wrote: > Could it be a consequence of how REPEATABLE READ transactions handle > snapshots? With REPEATABLE READ the snapshot is acquired only once at the > beginning of a transaction; a READ COMMITTED transaction re-evaluates