> Hi,
>
> I would like to propose to add an option to pgbench so that benchmark
> can quit immediately when any client is aborted. Currently, when a
> client is aborted due to some error, for example, network trouble,
> other clients continue their run until a certain number of transactions
> spe
On Sat, Aug 5, 2023 at 1:39 PM Thomas Munro wrote:
> Here is a rebase over 26669757, which introduced
> PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT.
Oops, please disregard v7 (somehow lost a precious line of code). V8 is better.
From 6544931e533aa015f39215f9c9d2df3e06700a96 Mon Sep 17 00:00:00 2001
Fr
On Sat, Aug 5, 2023 at 12:43 AM Christoph Berg wrote:
> I managed to reproduce it on the shell by running the test in a loop a
> few times. The failure looks like this:
It's great that you can reproduce this semi-reliably! I've rebased
the patch, hoping you can try it out.
https://www.postgresq
Here is a rebase over 26669757, which introduced
PROCSIG_RECOVERY_CONFLICT_LOGICALSLOT.
I got a bit confused about why this new conflict reason didn't follow
the usual ERROR->FATAL promotion rules and pinged Andres who provided:
"Logical decoding slots are only acquired while performing logical
d
On Fri, Aug 4, 2023 at 4:34 PM Tomas Vondra
wrote:
> Thanks for the clarification, I think I understand better now.
Honestly, it's gratifying to be understood at all in a discussion like
this one. Figuring out how to articulate some of my more subtle ideas
(without leaving out important nuances)
On 8/4/23 04:07, Peter Geoghegan wrote:
> On Thu, Aug 3, 2023 at 3:04 PM Tomas Vondra
> wrote:
>> Because my patch is all about reducing the heap pages, which are usually
>> the expensive part of the index scan. But you're right the "index scan"
>> with index filter may access more index pages, be
On 2023-08-03 23:55, Andy Fan wrote:
> The double quotes look weird to me. but it looks like a common
> situation.
>
> select numeric('1'::int); -- failed.
> select "numeric"('1'::int); -- ok.
It arises when you have an object (type, function, cast, whatever) whose
name in the catalog is the sa
Hi hackers,
Melih Mutlu , 16 Haz 2023 Cum, 17:03 tarihinde şunu
yazdı:
> With this change, here's a query to find how much space used by each
> context including its children:
>
> > WITH RECURSIVE cte AS (
> > SELECT id, total_bytes, id as root, name as root_name
> > FROM memory_contexts
On 2023-Jul-28, Alvaro Herrera wrote:
> To avoid that, one option would be to make this NN constraint
> undroppable ... but I don't see how. One option might be to add a
> pg_depend row that links the NOT NULL constraint to its PK constraint.
> But this will be a strange case that occurs nowhere
On Fri, Aug 4, 2023 at 4:47 AM Tomas Vondra
wrote:
> Well, I presented multiple options, so "yes" doesn't really clarify
> which of them applies. But my understanding is you meant the index pages
> accesses.
Sorry. Your understanding of what I must have meant before was correct
-- your patch pick
I have pushed Release 17 of the PostgreSQL Buildfarm client.
Release 17 has two main features:
* Modernize the way we do cross-version upgrade tests. Most of the
logic for modifying instances to make them suitable for cross
version upgrade testing has now been migrated to the Postgres cor
"Drouvot, Bertrand" writes:
> Now that fa88928470 generates automatically code and documentation
> related to wait events, why not exposing the wait events description
> through a system catalog relation?
I think you'd be better off making this a view over a set-returning
function. The nearby wo
Hi hackers,
while working on the new system catalog pg_wait_event (see [1]) I noticed that
some wait
events are currently "duplicated":
postgres=# select wait_event_name,count(*) from pg_wait_event group by
wait_event_name having count(*) > 1;
wait_event_name | count
-+---
Hi hackers,
Now that fa88928470 generates automatically code and documentation
related to wait events, why not exposing the wait events description
through a system catalog relation? (the idea has been proposed on twitter
by Yves Colin [1]).
I think that could be useful to:
- join this new rela
On Fri, 4 Aug 2023 at 03:44, Alvaro Herrera wrote:
> On 2023-Aug-03, Dave Cramer wrote:
>
> > New patch attached which uses PREPARED_SUB_COMMAND and
> > PORTAL_SUB_COMMAND instead
>
> Hmm, I would keep the prefix in this case and make the message type a
> second prefix, with the subtype last -
Dear Amit,
> So, we have three options here (a) As you have done in the patch,
> document this limitation and request user to perform some manual steps
> to drop the subscription; (b) don't allow upgrade to proceed if there
> are invalid slots in the old cluster; (c) provide a new function like
>
Re: Noah Misch
> On Tue, Jul 25, 2023 at 01:56:41PM +0530, Bharath Rupireddy wrote:
> > I've observed the following failure once in one of my Cirrus CI runs
> > on Windows Server on HEAD:
> >
> > timed out waiting for match: (?^:User was holding shared buffer pin
> > for too long) at
> > C:/cirrus
On 8/4/23 02:08, Peter Geoghegan wrote:
> On Thu, Aug 3, 2023 at 3:04 PM Tomas Vondra
> wrote:
>> When you say "index page accesses" do you mean accesses to index pages,
>> or accesses to heap pages from the index scan?
>
> Yes, that's exactly what I mean. Note that that's the dominant cost
>
On Fri, Aug 4, 2023 at 2:44 PM Drouvot, Bertrand
wrote:
>
> Hi,
>
> On 7/28/23 4:39 PM, Bharath Rupireddy wrote:
> > On Mon, Jul 24, 2023 at 9:00 AM Amit Kapila wrote:
> >>
> >>> 2. All candidate standbys will start one slot sync worker per logical
> >>> slot which might not be scalable.
> >>
> >
On Wed, Aug 2, 2023 at 1:43 PM Hayato Kuroda (Fujitsu)
wrote:
>
> > 3.
> > + /*
> > + * Get replication slots.
> > + *
> > + * XXX: Which information must be extracted from old node? Currently three
> > + * attributes are extracted because they are used by
> > + * pg_create_logical_replication_slo
> On Thu, Aug 03, 2023 at 12:23:52PM -0500, Tristan Partin wrote:
>
> This is the first I am learning about clang plugins. Interesting concept.
> Did you give any thought to using libclang instead of a compiler plugin? I
> am kind of doing similar work, but I went with libclang instead of a plugin.
In get_memoize_path() we have a theory about avoiding memoize node if
there are volatile functions in the inner rel's target/restrict list.
/*
* We can't use a memoize node if there are volatile functions in the
* inner rel's target list or restrict list. A cache hit could reduce the
* num
Oversight the DetachPartitionFinalize() again, I found the root cause why
'r_p_id_fkey' wat not removed.
DetachPartitionFinalize() call the GetParentedForeignKeyRefs() func to get
tuple from pg_constraint that will be delete but failed.
according to the comments, the GetParentedForeignKeyRefs() f
On Wed, Aug 2, 2023 at 8:20 AM Amit Kapila wrote:
>
> On Tue, Aug 1, 2023 at 12:11 PM Alvaro Herrera
> wrote:
> >
> > On 2023-Aug-01, Peter Smith wrote:
> >
> > > PSA a small patch making those above-suggested changes. The 'make
> > > check' and TAP subscription tests are all passing OK.
> >
> >
Hi,
On Fri, Aug 4, 2023 at 19:01 Erik Rijkers wrote:
> Op 7/21/23 om 12:33 schreef Amit Langote:
> >
> > Thanks for taking a look.
> >
>
> Hi Amit,
>
> Is there any chance to rebase the outstanding SQL/JSON patches, (esp.
> json_query)?
Yes, working on it. Will post a WIP shortly.
> --
Thank
Op 7/21/23 om 12:33 schreef Amit Langote:
Thanks for taking a look.
Hi Amit,
Is there any chance to rebase the outstanding SQL/JSON patches, (esp.
json_query)?
Thanks!
Erik Rijkers
On 2023-Aug-03, Dave Cramer wrote:
> New patch attached which uses PREPARED_SUB_COMMAND and
> PORTAL_SUB_COMMAND instead
Hmm, I would keep the prefix in this case and make the message type a
second prefix, with the subtype last -- PQMSG_CLOSE_PREPARED,
PQMSG_DESCRIBE_PORTAL and so on.
You def
Hi,
On 7/10/23 11:14 AM, Daniel Gustafsson wrote:
On 4 Apr 2023, at 12:04, Drouvot, Bertrand wrote:
On 4/3/23 11:47 PM, Gregory Stark (as CFM) wrote:
On Thu, 16 Mar 2023 at 05:25, Drouvot, Bertrand
wrote:
My plan was to get [1] done before resuming working on the "Split index and table
sta
Hi,
On 7/28/23 4:39 PM, Bharath Rupireddy wrote:
On Mon, Jul 24, 2023 at 9:00 AM Amit Kapila wrote:
2. All candidate standbys will start one slot sync worker per logical
slot which might not be scalable.
Yeah, that doesn't sound like a good idea but IIRC, the proposed patch
is using one wo
Oversight the DetachPartitionFinalize(), I found another bug below:
postgres=# CREATE TABLE p ( id bigint PRIMARY KEY ) PARTITION BY list (id);
CREATE TABLE
postgres=# CREATE TABLE p_1 PARTITION OF p FOR VALUES IN (1);
CREATE TABLE
postgres=# CREATE TABLE r_1 (
postgres(# id bigint PRIMARY K
On Fri, Jul 28, 2023 at 7:28 PM Ashutosh Bapat
wrote:
> >
> > + bms_free(child_sjinfo->commute_above_l);
> > + bms_free(child_sjinfo->commute_above_r);
> > + bms_free(child_sjinfo->commute_below_l);
> > + bms_free(child_sjinfo->commute_below_r);
> >
> > These four members in SpecialJoinIn
Le jeu. 3 août 2023 à 15:17, Christoph Moench-Tegeder
a écrit :
> ## Laetitia Avrot (laetitia.av...@gmail.com):
>
> > For my customer, their use case is to be able from an SQL client to
> double
> > check they're on the right host before doing things that could become a
> > production disaster.
Thank you for the feedback received so far. Sorry, I have become busy
lately with other work.
IIUC there is a general +1 for the idea, but I still have some
juggling necessary to make the functions/macros of patch 0001
acceptable to everybody.
The difficulties arise from:
- no function overloadin
Hi,
On 7/24/23 4:32 AM, Bharath Rupireddy wrote:
On Fri, Jul 21, 2023 at 5:16 PM shveta malik wrote:
Here are my thoughts about this feature:
Thanks for looking at it!
Important considerations:
1. Does this design guarantee the row versions required by subscribers
aren't removed on ca
34 matches
Mail list logo
