Amit Kapila writes:
> Any other ideas?
Given that the crash is so far down inside __dlopen(), and that there's
a clear reference to the string we presumably passed to that:
#11 0x7f518485e489 in _dl_open (file=0x55b692f2d2b0
On Fri, Sep 29, 2017 at 9:12 PM, Robert Haas wrote:
>
> It's possible that we might find that neither of the above approaches
> are practical and that the performance benefits of resolving the
> transaction from the original connection are large enough that we want
> to try
On Tue, Oct 3, 2017 at 8:31 AM, Amit Kapila wrote:
> On Tue, Oct 3, 2017 at 3:04 AM, Andreas Seltenreich
> wrote:
>> Hi,
>>
>> doing low-memory testing with REL_10_STABLE at 1f19550a87 also produced
>> a couple of parallel worker core dumps with the
On Thu, Sep 28, 2017 at 8:57 PM, Alexander Kuzmenkov
wrote:
> Hi Ashutosh,
>
> Thanks for the review.
>
> Jeff, I'm copying you because this is relevant to our discussion about what
> to do with mergeopfamilies when adding new merge join types.
>
> You have renamed
> Tatsuo Ishii writes:
>> I saw this while restoring 9.6 database to 10.0 database.
>> \connect: FATAL: conversion between UTF8 and MULE_INTERNAL is not supported
>> Is this expected? I saw nothing releated to this in the release notes.
>
> Don't think that's ever been
On Tue, Oct 3, 2017 at 1:15 PM, Zeus Kronion wrote:
> I previously made one minuscule contribution to the project two years ago.
> I'm interested in doing some more, and I'm trying to figure out what to
> focus on. Two SSL-related projects caught my attention:
> 1) Allow
Zeus Kronion writes:
> 2) I was surprised to learn the following from the docs:
>> By default, PostgreSQL will not perform any verification of the server
>> certificate.
> Is there a technical reason to perform no verification by default? Wouldn't
> a safer default be
Tatsuo Ishii writes:
> I saw this while restoring 9.6 database to 10.0 database.
> \connect: FATAL: conversion between UTF8 and MULE_INTERNAL is not supported
> Is this expected? I saw nothing releated to this in the release notes.
Don't think that's ever been supported.
> I saw this while restoring 9.6 database to 10.0 database.
>
> \connect: FATAL: conversion between UTF8 and MULE_INTERNAL is not supported
>
> Is this expected? I saw nothing releated to this in the release notes.
This had been allowed in 9.6. So I think 10.0 silently drops the feature.
Best
At Tue, 3 Oct 2017 10:23:08 +0900, Michael Paquier
wrote in
> On Tue, Oct 3, 2017 at 12:01 AM, Stephen Frost wrote:
> > I certainly don't care for the idea of adding log messages
I previously made one minuscule contribution to the project two years ago.
I'm interested in doing some more, and I'm trying to figure out what to
focus on. Two SSL-related projects caught my attention:
1) Allow automatic selection of SSL client certificates from a certificate
store
Hi David.
Thanks a lot for your review comments and sorry it took me a while to reply.
On 2017/09/28 18:16, David Rowley wrote:
> On 27 September 2017 at 14:22, Amit Langote
> wrote:
>> - 0001 includes refactoring that Dilip proposed upthread [1] (added him as
>>
Andres Freund writes:
> On 2017-10-02 19:50:51 -0400, Tom Lane wrote:
>> What I saw was that the backend process was consuming 100% of (one) CPU,
>> while the I/O transaction rate viewed by "iostat 1" started pretty low
>> --- under 10% of what the machine is capable of ---
On Tue, Oct 3, 2017 at 3:04 AM, Andreas Seltenreich wrote:
> Hi,
>
> doing low-memory testing with REL_10_STABLE at 1f19550a87 also produced
> a couple of parallel worker core dumps with the backtrace below.
> Although most of the backtrace is inside the dynamic linker, it
On Sun, Oct 1, 2017 at 5:18 AM, Nikolay Shaplov wrote:
> src/backend/access/common/reloptions.c get only 7 lines, it was quite covered
> by existing test, but all most of the access methods gets some coverage
> increase:
>
> src/backend/access/brin 1268 -> 1280 (+18)
>
On Thu, Sep 21, 2017 at 8:21 AM, Ashutosh Bapat
wrote:
> Here's set of rebased patches. The patch with extra tests is not for
> committing. All other patches, except the last one, will need to be
> committed together. The last patch may be committed along with
On Tue, Oct 3, 2017 at 12:54 AM, Alvaro Herrera wrote:
> Michael Paquier wrote:
>> On Mon, Oct 2, 2017 at 11:44 PM, Robert Haas wrote:
>> > On Mon, Oct 2, 2017 at 10:19 AM, Michael Paquier
>> > wrote:
>> >> So those bits
Hi,
I'd recently noticed that expr* functions actually show up in profiles
because we use them at some very common paths
(e.g. ExecTypeFromTLInternal()) and that we commonly call all the three
variants from $subject in sequence.
Looking at their code I was wondering whether it's reasonable to
On Tue, Oct 3, 2017 at 1:30 AM, Robert Haas wrote:
> On Fri, Sep 15, 2017 at 6:29 PM, Michael Paquier
> wrote:
>> I would like to point out that per the RFC, if the client attempts a
>> SSL connection with SCRAM and that the server supports
Hi,
Sorry, I saw this once but somehow my attension was blown away on
the way.
At Tue, 3 Oct 2017 02:41:34 +0300, Alexander Korotkov
wrote in
> On Tue, Oct 3, 2017 at 12:20 AM, Maksim Milyutin
On Tue, Oct 3, 2017 at 3:12 AM, Robert Haas wrote:
> On Mon, Oct 2, 2017 at 11:57 AM, Tom Lane wrote:
> > Daniel Gustafsson writes:
> >> Thanks to everyone who participated, and to everyone who have responded
> to my
> >> nagging via
On Tue, Oct 3, 2017 at 12:01 AM, Stephen Frost wrote:
> I certainly don't care for the idea of adding log messages saying we
> aren't doing anything just to match a count that's incorrectly claiming
> that checkpoints are happening when they aren't.
>
> The down-thread
On Tue, Oct 3, 2017 at 9:07 AM, Alexander Korotkov
wrote:
> +1,
> I see 3 options there:
> 1) Drop high-order bit, as you proposed.
> 2) Allow negative queryIds.
> 3) Implement unsigned 64-type.
>
> #1 causes minor loss of precision which looks rather insignificant in
On Tue, Oct 3, 2017 at 6:54 AM, Tom Lane wrote:
> I wrote:
>> If this is the only problem then I'd agree we should stick to a spinlock
>> (I assume the strings in question can't be very long). I was thinking
>> more about what to do if we find other violations that are harder
I saw this while restoring 9.6 database to 10.0 database.
\connect: FATAL: conversion between UTF8 and MULE_INTERNAL is not supported
Is this expected? I saw nothing releated to this in the release notes.
Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English:
On 2017/10/03 7:16, Michael Paquier wrote:
> On Tue, Oct 3, 2017 at 1:23 AM, Alexander Korotkov
> wrote:
>> On Mon, Oct 2, 2017 at 6:57 PM, Tom Lane wrote:
>>>
>>> Daniel Gustafsson writes:
Thanks to everyone who participated,
On Tue, Oct 3, 2017 at 12:32 AM, Tom Lane wrote:
> Peter Geoghegan writes:
> > You need to change the SQL interface as well, although I'm not sure
> > exactly how. The problem is that you are now passing a uint64 queryId
> > to Int64GetDatumFast() within
On 2017-10-02 19:50:51 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On 2017-10-02 18:33:17 -0400, Tom Lane wrote:
> >> I'm kind of surprised that machine B doesn't show obvious tanking in this
> >> test given that msync() makes it suck so badly at copying a database.
>
Andres Freund writes:
> On 2017-10-02 18:33:17 -0400, Tom Lane wrote:
>> I'm kind of surprised that machine B doesn't show obvious tanking in this
>> test given that msync() makes it suck so badly at copying a database.
>> I wonder what is different from the kernel's
On Tue, Oct 3, 2017 at 12:20 AM, Maksim Milyutin
wrote:
> I have tested the following case:
>
> create type pair as (x int, y int);
> prepare test as select json_populate_record(null::pair, '{"x": 1, "y":
> 2}'::json);
> drop type pair cascade;
>
> execute test;
>
> -- The
Hi,
On 2017-10-02 18:33:17 -0400, Tom Lane wrote:
> Andres Freund writes:
> > To demonstrate what I'm observing here, on linux with a fairly fast ssd:
> > ...
>
> I tried to replicate this test as closely as I could on the Mac hardware
> I have laying about.
Thanks!
> I
On Mon, Oct 2, 2017 at 02:12:50PM -0400, Stephen Frost wrote:
> > How should this be handled for the Postgres 11 release notes?
>
> Ideally, we would let the individuals choose how to be recognized in
> release notes, and anywhere else we recognize them. We have the start
> of that in
Andres Freund writes:
> To demonstrate what I'm observing here, on linux with a fairly fast ssd:
> ...
I tried to replicate this test as closely as I could on the Mac hardware
I have laying about. I only bothered with the synchronous_commit=off
case, though, since you say
On Tue, Oct 3, 2017 at 1:23 AM, Alexander Korotkov
wrote:
> On Mon, Oct 2, 2017 at 6:57 PM, Tom Lane wrote:
>>
>> Daniel Gustafsson writes:
>> > Thanks to everyone who participated, and to everyone who have responded
>> > to my
>>
On 2017-10-02 17:57:51 -0400, Tom Lane wrote:
> Andres Freund writes:
> > Done that way. It's a bit annoying, because we've to take care to
> > initialize the "unused" part of the array with a valid signalling it's
> > an unused mapping. Can't use 0 for that because
Andres Freund writes:
> Done that way. It's a bit annoying, because we've to take care to
> initialize the "unused" part of the array with a valid signalling it's
> an unused mapping. Can't use 0 for that because fmgr_builtins[0] is a
> valid entry.
The prototype code I
I wrote:
> If this is the only problem then I'd agree we should stick to a spinlock
> (I assume the strings in question can't be very long). I was thinking
> more about what to do if we find other violations that are harder to fix.
I took a quick look through walreceiver.c, and there are a
On 2017-09-28 19:06:27 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On 2017-09-28 18:52:28 -0400, Tom Lane wrote:
> >> Uh, what? Access to fmgr_nbuiltins shouldn't be part of any critical path
> >> anymore after this change.
>
> > Indeed. But the size of the the oid ->
Hi,
doing low-memory testing with REL_10_STABLE at 1f19550a87 also produced
a couple of parallel worker core dumps with the backtrace below.
Although most of the backtrace is inside the dynamic linker, it looks
like it was passed a pointer to gone-away shared memory.
regards,
Andreas
Core was
Andres Freund writes:
> On 2017-10-02 17:30:25 -0400, Tom Lane wrote:
>> Or replace the spinlock with an LWLock?
> That'd probably be a good idea, but I'm loathe to do so in the back
> branches. Not at this callsite, but some others, there's some potential
> for contention.
Peter Geoghegan writes:
> You need to change the SQL interface as well, although I'm not sure
> exactly how. The problem is that you are now passing a uint64 queryId
> to Int64GetDatumFast() within pg_stat_statements_internal(). That
> worked when queryId was a uint32, because you
On 2017-10-02 17:30:25 -0400, Tom Lane wrote:
> Andres Freund writes:
> > Yes, that'd be a bad idea. It's not great to have memcpys in a critical
> > section, but it's way better than pallocs. So we need to use some local
> > buffers that this get copied to.
>
> Or replace
Andres Freund writes:
> On 2017-10-02 22:56:49 +0200, Andreas Seltenreich wrote:
>> low-memory testing with REL_10_STABLE at 1f19550a87 produced the
>> following PANIC:
>> stuck spinlock detected at pg_stat_get_wal_receiver, walreceiver.c:1397
> Ugh.
Egad.
> Yes, that'd be
Hi, Alexander!
Thanks for the comments.
02.10.17 20:02, Alexander Korotkov wrote:
Please, register this patch at upcoming commitfest to ensure it
wouldn't be forgotten.
Regression test changes (both .sql and .out) are essential parts of
the patch. I see no point in posting them separately.
On Mon, Oct 2, 2017 at 9:10 AM, Robert Haas wrote:
> On Mon, Oct 2, 2017 at 11:02 AM, Joshua D. Drake
> wrote:
>> +1 to both of these as well.
>
> OK, so here's a patch. Review appreciated.
You need to change the SQL interface as well, although
> On Wed, Sep 27, 2017 at 8:29 PM, Michael Paquier
> wrote:
>> Looking at this list, the first name is followed by the family name,
>> so there are inconsistencies with some Japanese names:
>> - Fujii Masao should be Masao Fujii.
>> - KaiGai Kohei should be Kohei
On 2017-10-02 22:56:49 +0200, Andreas Seltenreich wrote:
> Hi,
>
> low-memory testing with REL_10_STABLE at 1f19550a87 produced the
> following PANIC:
>
> stuck spinlock detected at pg_stat_get_wal_receiver, walreceiver.c:1397
Ugh.
> I was about to wrap the pstrdup()s with a PG_TRY block,
On Tue, Oct 3, 2017 at 9:56 AM, Andreas Seltenreich wrote:
> low-memory testing with REL_10_STABLE at 1f19550a87 produced the
> following PANIC:
>
> stuck spinlock detected at pg_stat_get_wal_receiver, walreceiver.c:1397
>
> I was about to wrap the pstrdup()s with a PG_TRY
Hi,
low-memory testing with REL_10_STABLE at 1f19550a87 produced the
following PANIC:
stuck spinlock detected at pg_stat_get_wal_receiver, walreceiver.c:1397
I was about to wrap the pstrdup()s with a PG_TRY block, but I can't find
a spinlock being released in a PG_CATCH block anywhere, so
On 03/10/17 04:02, Joshua D. Drake wrote:
On 10/01/2017 04:22 PM, Robert Haas wrote:
On Sun, Oct 1, 2017 at 3:48 PM, Greg Stark wrote:
Well these kinds of monitoring systems tend to be used by operations
people who are a lot more practical and a lot less worried about
On 2017-10-02 15:59:05 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On 2017-10-02 15:54:43 -0400, Tom Lane wrote:
> >> Should I expect there to be any difference at all? We don't enable
> >> *_flush_after by default on non-Linux platforms.
>
> > Right, you'd have to
Andres Freund writes:
> On 2017-10-02 15:54:43 -0400, Tom Lane wrote:
>> Should I expect there to be any difference at all? We don't enable
>> *_flush_after by default on non-Linux platforms.
> Right, you'd have to enable that. But your patch would neuter an
> intentionally
On 2017-10-02 15:54:43 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On 2017-10-02 15:42:25 -0400, Tom Lane wrote:
> >> I experimented with this further by seeing whether the msync() code path
> >> is of any value on Sierra either. The answer seems to be "no": cloning
>
Andres Freund writes:
> On 2017-10-02 15:42:25 -0400, Tom Lane wrote:
>> I experimented with this further by seeing whether the msync() code path
>> is of any value on Sierra either. The answer seems to be "no": cloning
>> a scale-1000 pgbench database takes about 17-18
On 2017-10-02 15:42:25 -0400, Tom Lane wrote:
> I wrote:
> > In short, therefore, APFS cannot cope with the way we're using msync().
>
> I experimented with this further by seeing whether the msync() code path
> is of any value on Sierra either. The answer seems to be "no": cloning
> a
Thanks for this breakdown Tom!
FWIW - I'm on Postgres 9.6.5 as bundled with Postgres.app (2.0.5) running
on 2013 MBP (2.7GHz i7 / 16GB / SSD) setup. It looks like this might be a
priority for an upcoming release, so I might try to hold out for
downstream, but thanks for the patch. It will help if
On Mon, Oct 2, 2017 at 12:44 PM, Tom Lane wrote:
>> And I'm saying - that argument is bogus. Regardless of what people
>> want or what we have historically done in the case where the
>> record/row is the only INTO target, when there are multiple targets it
>> seems clear that
I wrote:
> In short, therefore, APFS cannot cope with the way we're using msync().
I experimented with this further by seeing whether the msync() code path
is of any value on Sierra either. The answer seems to be "no": cloning
a scale-1000 pgbench database takes about 17-18 seconds on my Sierra
So yes, distinguishing stored vs. not stored computed columns is useful,
especially if the expression can refer to other columns of the same row,
though not only then.
Examples:
-- useful only if stored (assuming these never get updated)
inserted_at TIMESTAMP WITHOUT TIME ZONE AS
On Mon, Oct 2, 2017 at 9:11 PM, Andrew Borodin
wrote:
> On Mon, Oct 2, 2017 at 8:00 PM, Alexander Korotkov
> wrote:
> > What happen if exactly this "continue" fires?
> >
> >> if (GistFollowRight(stack->page))
> >> {
> >> if (!xlocked)
> >> {
>
On Mon, Oct 02, 2017 at 02:30:38PM -0400, Tom Lane wrote:
> Nico Williams writes:
> > On Mon, Oct 02, 2017 at 12:50:14PM -0400, Adam Brusselback wrote:
> >> So for me, i'd rather default to compute on read, as long storing the
> >> pre-computed value is an option when
On Mon, Oct 02, 2017 at 02:30:38PM -0400, Tom Lane wrote:
> Nico Williams writes:
> > On Mon, Oct 02, 2017 at 12:50:14PM -0400, Adam Brusselback wrote:
> >> So for me, i'd rather default to compute on read, as long storing the
> >> pre-computed value is an option when
Nico Williams writes:
> On Mon, Oct 02, 2017 at 12:50:14PM -0400, Adam Brusselback wrote:
>> So for me, i'd rather default to compute on read, as long storing the
>> pre-computed value is an option when necessary.
> Sure, I agree. I was just wondering whether there might
Brent Dearth writes:
> I just recently "upgraded" to High Sierra and experiencing horrendous CREATE
> DATABASE performance. Creating a database from a 3G template DB used to
> take ~1m but post-upgrade is taking ~22m at a sustained write of around
> 4MB/s. Occasionally,
Bruce,
* Bruce Momjian (br...@momjian.us) wrote:
> On Fri, Sep 29, 2017 at 12:00:05PM -0400, Peter Eisentraut wrote:
> > On 9/29/17 11:35, Robert Haas wrote:
> > > On Wed, Sep 27, 2017 at 8:29 PM, Michael Paquier
> > > wrote:
> > >> Looking at this list, the first name
On Mon, Oct 2, 2017 at 8:00 PM, Alexander Korotkov
wrote:
> What happen if exactly this "continue" fires?
>
>> if (GistFollowRight(stack->page))
>> {
>> if (!xlocked)
>> {
>> LockBuffer(stack->buffer, GIST_UNLOCK);
>> LockBuffer(stack->buffer, GIST_EXCLUSIVE);
>>
On Mon, Oct 02, 2017 at 12:50:14PM -0400, Adam Brusselback wrote:
> I know that for my use-cases, having both options available would be very
> appreciated. The vast majority of the computed columns I would use in my
> database would be okay to compute on read. But there are for sure some
>
On Fri, Sep 29, 2017 at 12:00:05PM -0400, Peter Eisentraut wrote:
> On 9/29/17 11:35, Robert Haas wrote:
> > On Wed, Sep 27, 2017 at 8:29 PM, Michael Paquier
> > wrote:
> >> Looking at this list, the first name is followed by the family name,
> >> so there are
On Mon, Oct 2, 2017 at 9:42 AM, Peter Eisentraut
wrote:
> I understand where you are coming from. My experience with developing
> this feature has been that it is quite fragile in some ways. We have
> had this out there for testing for many months, and we have
I have faced two issues with logical replication.
Repro scenario:
1. start two Postgres instances (I start both at local machine).
2. Initialize pgbench tables at both instances:
pgbench -i -s 10 postgres
3. Create publication of pgbench_accounts table at one node:
create publication pub
2017-10-02 18:44 GMT+02:00 Tom Lane :
> Robert Haas writes:
> > On Mon, Oct 2, 2017 at 12:28 PM, Tom Lane wrote:
> >> I'm not sure if that's true or not. I am sure, though, that since
> >> we've done B for twenty years we can't
On Mon, Oct 2, 2017 at 3:43 PM, Maksim Milyutin
wrote:
> On 26.09.2017 23:25, Maksim Milyutin wrote:
>
> Updated patchset contains more transparent definition of composite type
> for constant node and regression test for described above buggy case.
>
> Is there any interest
On 02/10/17 17:49, Konstantin Knizhnik wrote:
> I have faced two issues with logical replication.
> Reproducing scenario:
>
> 1. start two Postgres instances (I start both at local machine).
> 2. Initialize pgbench tables at both instances:
> pgbench -i -s 10 postgres
> 3. Create publication
On Sun, Oct 1, 2017 at 11:53 AM, Shubham Barai
wrote:
> Yes, page-level predicate locking should happen only when fast update is
> off.
> Actually, I forgot to put conditions in updated patch. Does everything
> else look ok ?
>
I think that isolation tests should be
I know that for my use-cases, having both options available would be very
appreciated. The vast majority of the computed columns I would use in my
database would be okay to compute on read. But there are for sure some
which would be performance prohibitive to have compute on read, so i'd
rather
Robert Haas writes:
> On Mon, Oct 2, 2017 at 12:28 PM, Tom Lane wrote:
>> I'm not sure if that's true or not. I am sure, though, that since
>> we've done B for twenty years we can't just summarily change to A.
> I agree, but so what? You said that we
On 10/1/17 14:26, Peter Geoghegan wrote:
> It does seem too late. It's disappointing that we did not do better
> here. This problem was entirely avoidable.
I understand where you are coming from. My experience with developing
this feature has been that it is quite fragile in some ways. We have
On Mon, Oct 2, 2017 at 12:28 PM, Tom Lane wrote:
> Robert Haas writes:
>> On Tue, Sep 19, 2017 at 3:18 PM, Tom Lane wrote:
>>> That's certainly a case that we ought to support somehow. The problem is
>>> staying reasonably
On Fri, Sep 15, 2017 at 6:29 PM, Michael Paquier
wrote:
> I would like to point out that per the RFC, if the client attempts a
> SSL connection with SCRAM and that the server supports channel
> binding, then it has to publish the SASL mechanism for channel
> binding,
Robert Haas writes:
> On Tue, Sep 19, 2017 at 3:18 PM, Tom Lane wrote:
>> That's certainly a case that we ought to support somehow. The problem is
>> staying reasonably consistent with the two-decades-old precedent of the
>> existing behavior for one
On Mon, Oct 2, 2017 at 6:57 PM, Tom Lane wrote:
> Daniel Gustafsson writes:
> > Thanks to everyone who participated, and to everyone who have responded
> to my
> > nagging via the CF app email function. This is clearly an awesome
> community.
>
> And thanks
On Tue, Sep 19, 2017 at 3:18 PM, Tom Lane wrote:
> Robert Haas writes:
>> I think the fact that single-target INTO lists and multiple-target
>> INTO lists are handled completely differently is extremely poor
>> language design. It would have been far
On 2017-10-02 23:24:30,"Alexander Korotkov" wrote:
On Sun, Oct 1, 2017 at 8:27 PM, chenhj wrote:
Now, this patch looks good for me. It applies cleanly, builds cleanly, passes
regression tests, new functionality is covered by regression tests.
On Mon, Oct 2, 2017 at 11:05 AM, Michael Paquier
wrote:
> Well, there are cases where you don't need any locking checks, and the
> proposed patch ignores that.
I understand that, but shouldn't we then look for a way to adjust the
patch so that it doesn't have that
On Mon, Oct 2, 2017 at 11:57 AM, Tom Lane wrote:
> Daniel Gustafsson writes:
>> Thanks to everyone who participated, and to everyone who have responded to my
>> nagging via the CF app email function. This is clearly an awesome community.
>
> And thanks to you
On Mon, Oct 2, 2017 at 11:02 AM, Joshua D. Drake wrote:
> +1 to both of these as well.
OK, so here's a patch. Review appreciated.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
64-bit-queryid-v1.patch
Description: Binary
Daniel Gustafsson writes:
> Thanks to everyone who participated, and to everyone who have responded to my
> nagging via the CF app email function. This is clearly an awesome community.
And thanks to you for your hard work as CFM! That's tedious and
largely thankless work, but
Michael Paquier wrote:
> On Mon, Oct 2, 2017 at 11:44 PM, Robert Haas wrote:
> > On Mon, Oct 2, 2017 at 10:19 AM, Michael Paquier
> > wrote:
> >> So those bits could be considered for integration.
> >
> > Yes, and they also tend to suggest that
I just recently "upgraded" to High Sierra and experiencing horrendous CREATE
DATABASE performance. Creating a database from a 3G template DB used to
take ~1m but post-upgrade is taking ~22m at a sustained write of around
4MB/s. Occasionally, attempting to create an empty database hangs
I have faced two issues with logical replication.
Reproducing scenario:
1. start two Postgres instances (I start both at local machine).
2. Initialize pgbench tables at both instances:
pgbench -i -s 10 postgres
3. Create publication of pgbench_accounts table at one node:
create
We have now entered October, which marks the end of Commitfest 201709. All
patches have now been dealt with and the commitfest is closed. Before starting
on moving, and closing, patches the stat for this commitfest were:
Needs review: 69.
Waiting on Author: 22.
Ready for
On Sun, Oct 1, 2017 at 8:27 PM, chenhj wrote:
> On 2017-10-01 04:09:19,"Alexander Korotkov"
> wrote:
>
> On Sat, Sep 30, 2017 at 8:18 PM, chenhj wrote:
>
>> On 2017-09-30 02:17:54,"Alexander Korotkov" >
On Thu, Aug 31, 2017 at 12:16:43AM -0400, Peter Eisentraut wrote:
> In previous discussions, it has often been a source of confusion whether
> these generated columns are supposed to be computed on insert/update and
> stored, or computed when read. The SQL standard is not explicit, but
> appears
On Mon, Oct 2, 2017 at 11:44 PM, Robert Haas wrote:
> On Mon, Oct 2, 2017 at 10:19 AM, Michael Paquier
> wrote:
>> So those bits could be considered for integration.
>
> Yes, and they also tend to suggest that the rest of the patch has value.
On 10/01/2017 04:22 PM, Robert Haas wrote:
On Sun, Oct 1, 2017 at 3:48 PM, Greg Stark wrote:
Well these kinds of monitoring systems tend to be used by operations
people who are a lot more practical and a lot less worried about
theoretical concerns like that.
+1, well said.
Vik, all,
* Vik Fearing (vik.fear...@2ndquadrant.com) wrote:
> I recently had a sad because I noticed that checkpoint counts were
> increasing in pg_stat_bgwriter, but weren't accounted for in my logs
> with log_checkpoints enabled.
> After some searching, I found that it was the idle
Hi, Andrew!
On Mon, Oct 2, 2017 at 1:40 PM, Andrew Borodin
wrote:
> Thanks for looking into the patch!
>
> On Thu, Sep 28, 2017 at 3:59 PM, Alexander Korotkov <
> a.korot...@postgrespro.ru> wrote:
>
>>
>>
>> In gistdoinsert() you do CheckForSerializableConflictIn() only
On Mon, Oct 2, 2017 at 10:19 AM, Michael Paquier
wrote:
> So those bits could be considered for integration.
Yes, and they also tend to suggest that the rest of the patch has value.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL
On Thu, Sep 14, 2017 at 10:45 PM, Amit Kapila wrote:
> The latest patch again needs to be rebased. Find rebased patch
> attached with this email.
I read through this patch this morning. Copying Tom in the hopes
that he might chime in on the following two issues in
On Mon, Oct 2, 2017 at 9:09 PM, Robert Haas wrote:
> I think the first question we ought to be asking ourselves is whether
> any of the PageGetLSN -> BufferGetLSNAtomic changes the patch
> introduces are live bugs. If they are, then we ought to fix those
> separately (and
1 - 100 of 132 matches
Mail list logo