Hi!
> 27 дек. 2018 г., в 12:54, Evgeniy Efimkin написал(а):
>
> In latest patch i removed `FOR ALL TABLES` clause and `alltables` parameter,
> now it's look more simple.
> Add new system view pg_user_subscirption to allow non-superuser use pg_dump
> and select addition column from pg_subscrpti
On Sun, 13 Jan 2019 at 00:04, Tomas Vondra wrote:
> On 1/12/19 8:49 AM, Dean Rasheed wrote:
> > A possible refinement would be to say that if there are more than
> > stats_target items more common than this mincount threshold, rather than
> > excluding the least common ones to get the target numbe
(Removing Adrien from the CC list, because messages to that address
keep bouncing)
On Sun, 13 Jan 2019 at 00:31, Tomas Vondra wrote:
>
> On 1/10/19 6:09 PM, Dean Rasheed wrote:
> >
> > In the previous discussion around UpdateStatisticsForTypeChange(), the
> > consensus appeared to be that we shou
On 12/01/2019 03:12, Andres Freund wrote:
On 2019-01-10 15:58:41 -0800, Andres Freund wrote:
A number of postgres files have sections like heapam's
* INTERFACE ROUTINES
...
They're often out-of-date, and I personally never found them to be
useful. A few people, including yours truly, have bee
Peter Eisentraut wrote:
> Here is an updated patch.
On a table with pre-existing contents, the creation of a unique index
does not seem to detect the duplicates that are equal per the
collation and different binary-wise.
postgres=# \d test3ci
Table "public.test3ci"
Colum
On 30/12/2018 22:56, Tom Lane wrote:
This comment needs copy-editing:
+ * If the user hits interrupts the startup (e.g. with CTRL-C), we'd
Looks good otherwise.
Thanks, fixed that, and pushed.
- Heikki
Thanks for the patch updates.
A few comments so far from me :
+static void _selectTableAccessMethod(ArchiveHandle *AH, const char
*tablespace);
tablespace => tableam
+_selectTableAccessMethod(ArchiveHandle *AH, const char *tableam)
+{
+ PQExpBuffer cmd = createPQExpBuffer();
createPQExpBuf
On Sun, Jan 13, 2019 at 3:06 PM Tom Lane wrote:
>
> There's still a logical problem here, which is that in order to
> prove that the ScalarArrayOpExpr yields NULL when its LHS does,
> you also have to prove that the RHS is not an empty array.
>
> Otherwise you're up against the fact that the OR of
On 1/10/19 8:44 AM, Peter Eisentraut wrote:
On 09/01/2019 19:49, Andreas Karlsson wrote:
Maybe this is orthogonal and best handled elsewhere but have you when
working with string equality given unicode normalization forms[1] any
thought?
Nondeterministic collations do address this by allowing
On 2019-Jan-13, Andres Freund wrote:
> On 2019-01-13 23:54:58 -0300, Alvaro Herrera wrote:
> > On 2019-Jan-13, Andres Freund wrote:
> >
> > > Alvaro, you'd introduced the split of HeapScanDesc and HeapScanDescData
> > > being in different files (in a3540b0f657c6352) - what do you think about
> >
On 2019-Jan-13, Andres Freund wrote:
> One split I am wondering about however is splitting out the sysstable_
> stuff out of genam.h. It's imo a different component and shouldn't
> really be in there. Would be quite a bit of rote work to add all the
> necessary includes over the backend...
Yeah -
On 1/14/19 12:20 PM, Dean Rasheed wrote:
> (Removing Adrien from the CC list, because messages to that address
> keep bouncing)
>
> On Sun, 13 Jan 2019 at 00:31, Tomas Vondra
> wrote:
>>
>> On 1/10/19 6:09 PM, Dean Rasheed wrote:
>>>
>>> In the previous discussion around UpdateStatisticsForTyp
On 1/11/19 12:26 AM, Peter Geoghegan wrote:
> On Thu, Jan 10, 2019 at 1:15 PM Robert Haas wrote:
>> I feel like there has been some other thread where this was discussed,
>> but I can't find it right now. I think that the "query construction"
>> logic in transformMergeStmt is fundamentally the
James Coleman writes:
> On Sun, Jan 13, 2019 at 3:06 PM Tom Lane wrote:
>> There's still a logical problem here, which is that in order to
>> prove that the ScalarArrayOpExpr yields NULL when its LHS does,
>> you also have to prove that the RHS is not an empty array.
> I've constructed a test ca
Andreas Karlsson wrote:
> > Nondeterministic collations do address this by allowing canonically
> > equivalent code point sequences to compare as equal. You still need a
> > collation implementation that actually does compare them as equal; ICU
> > does this, glibc does not AFAICT.
>
> A
On Mon, Jan 14, 2019 at 11:08 AM Tom Lane wrote:
>
> James Coleman writes:
> > On Sun, Jan 13, 2019 at 3:06 PM Tom Lane wrote:
> >> There's still a logical problem here, which is that in order to
> >> prove that the ScalarArrayOpExpr yields NULL when its LHS does,
> >> you also have to prove tha
James Coleman writes:
> On Mon, Jan 14, 2019 at 11:08 AM Tom Lane wrote:
>> I think these test cases don't actually prove much about the behavior
>> of your patch. Wouldn't they be handled by the generic OR-conversion
>> logic, since there's fewer than MAX_SAOP_ARRAY_SIZE items?
> Which ones ar
On Mon, Jan 14, 2019 at 11:34 AM Tom Lane wrote:
>
> James Coleman writes:
> > On Mon, Jan 14, 2019 at 11:08 AM Tom Lane wrote:
> >> I think these test cases don't actually prove much about the behavior
> >> of your patch. Wouldn't they be handled by the generic OR-conversion
> >> logic, since
James Coleman writes:
> Are those invariants we want to keep (and recognize as regression
> tests)? If so, I can confirm that they aren't duplicative of the rest
> of the file, and if not I can remove them.
I can't get terribly excited about them ... if we were changing their
behavior, or if ther
On Mon, 14 Jan 2019 at 15:45, Tomas Vondra
wrote:
>
>
> On 1/11/19 12:26 AM, Peter Geoghegan wrote:
> > On Thu, Jan 10, 2019 at 1:15 PM Robert Haas
> wrote:
> >> I feel like there has been some other thread where this was discussed,
> >> but I can't find it right now. I think that the "query co
ome of those rules. But
I'd be much happier if some of those asserts could be reinstated, even
if only in a weaker form.
regards
--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
0001-Add-logical_work_
0001 -- looks reasonable. One hunk in executor.h changes LockTupleMode
to "enum LockTupleMode", but there's no need for that.
AFAIK the only reason to have the struct FooBarData vs. FooBar (ptr)
split is so that it's possible to refer to structs without having
the full struct definition. I think
Hi,
Tom Lane writes:
> I'm afraid that this patch is likely to be, if not completely broken,
> at least very much less useful than one could wish by the time we get
> done closing the holes discussed in this other thread:
>
> https://www.postgresql.org/message-id/flat/5d910e2e-0db8-ec06-dd5f-baec
Hi,
On 2019-01-14 15:36:14 -0300, Alvaro Herrera wrote:
> 0001 -- looks reasonable. One hunk in executor.h changes LockTupleMode
> to "enum LockTupleMode", but there's no need for that.
Oh, that escaped from an earlier version where I briefly forgot that one
cannot portably forward-declare enums
Type() uses that, so
>> attribute_referenced will always be true.
> Hmmm. I'm pretty sure I came to the conclusion it's in fact necessary,
> but I might be wrong. Will check.
>
Turns out you were right - the attribute_referenced piece was quite
unnecessary. So I've removed it. I've also extended the regression tests
to verify changing type of another column does not reset the stats.
regards
--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
0001-multivariate-MCV-lists-20190114.patch.gz
Description: application/gzip
0002-multivariate-histograms-20190114.patch.gz
Description: application/gzip
David Rowley writes:
> I ran a few benchmarks on an AWS m5d.large instance based on top of
> c5c7fa261f5. The biggest regression I see is from a simple SELECT 1 at
> around 5-6%. A repeat of your test of SELECT 2+2 showed about half
> that regression so the simple addition function call is introdu
On 2019-Jan-14, Andres Freund wrote:
> I think the whole pointer hiding game that we play is a really really
> bad idea. We should just about *never* have typedefs that hide the fact
> that something is a pointer. But it's hard to go away from that for
> legacy reasons.
>
> The problem with your
While testing my nbtree heap TID keyspace patch, I came across a case
where amcheck reliably reports corruption. It appeared that a 4 byte
varlena index entry that was expected in an index was not actually
present. However, index scan queries with the "missing" value in their
qual didn't actually g
Hello Sergei,
> This patch correlates with my proposal
> "add session information column to pg_stat_statements"
> https://www.postgresql.org/message-id/3aa097d7-7c47-187b-5913-db8366cd4491%40gmail.com
> They both address the problem to identify the factors that make
> different execution plans fo
On 12/13/18 8:09 AM, Surafel Temesgen wrote:
>
>
> On Wed, Dec 12, 2018 at 9:28 PM Tomas Vondra
> mailto:tomas.von...@2ndquadrant.com>> wrote:
>
>
> Can you also update the docs to mention that the functions called from
> the WHERE clause does not see effects of the COPY itself?
>
>
>
Peter Geoghegan writes:
> The heapallindexed enhancement that made it into Postgres 11 assumes
> that the representation of index tuples produced by index_form_tuple()
> (or all relevant index_form_tuple() callers) is deterministic: for
> every possible heap tuple input there must be a single poss
On Mon, Jan 14, 2019 at 1:31 PM Tom Lane wrote:
> Peter Geoghegan writes:
> > The heapallindexed enhancement that made it into Postgres 11 assumes
> > that the representation of index tuples produced by index_form_tuple()
> > (or all relevant index_form_tuple() callers) is deterministic: for
> >
On 2019-Jan-14, Tomas Vondra wrote:
> The one slightly annoying issue is that currently all the options are
> formatted as text, including e.g. cpu_tuple_cost. That's because the GUC
> options may have show hook, so I can't access the value directly (not
> sure if there's an option around it).
I
Nikhil Sontakke writes:
> I'd like to believe that the latest patch set tries to address some
> (if not all) of your concerns. Can you please take a look and let me
> know?
Hi, sure.
General things:
- Earlier I said that there is no point of sending COMMIT PREPARED if
decoding snapshot bec
This patch is marked as "ready for committer", but that characterization
seems way over-optimistic. It looks like there are several unsettled
questions:
1. The connection parameter name and values are unlike the very similar
feature in pgJDBC. I think this is a fair complaint. Now I'm not in lo
Michael Paquier writes:
> On Wed, Sep 19, 2018 at 02:26:53PM +0200, Ildar Musin wrote:
>> At this point I'd like to ask community what in your opinion would be the
>> best course of action and whether this feature should be implemented within
>> libpq at all? Because from my POV there are factors
On 2019-Jan-09, Justin Pryzby wrote:
> -- Fails to error
> postgres=# CREATE UNIQUE INDEX ON t(j) INCLUDE(i);
>
> -- Fail to enforce uniqueness across partitions due to failure to enforce
> inclusion of partition key in index KEY
> postgres=# INSERT INTO t VALUES(1,1);
> postgres=# INSERT INTO t
On 2018-11-07 14:25:54 -0500, Tom Lane wrote:
> Robert Haas writes:
> > On Fri, May 11, 2018 at 11:01 AM, Simon Riggs wrote:
> I have no problem if you want to replace this with an even better
> design in a later release.
>
> >>> Meh. The author / committer should get a patch into the
On Mon, Jan 14, 2019 at 1:46 PM Peter Geoghegan wrote:
> I would have said that the assumption is that a fixed source tuple
> will generate identical index entries. The problem with that is that
> my idea of what constitutes a fixed input now seems to have been
> faulty. I didn't think that the ex
Tomas Vondra-4 wrote
> Hello Sergei,
>
>> This patch correlates with my proposal
>> "add session information column to pg_stat_statements"
>> https://www.postgresql.org/message-id/3aa097d7-7c47-187b-5913-db8366cd4491%40gmail.com
>> They both address the problem to identify the factors that make
>>
Andres Freund writes:
> On 2018-11-07 14:25:54 -0500, Tom Lane wrote:
>> In short, it seems likely to me that large parts of this patch need to
>> be pulled out, rewritten, and then put back in different places than
>> they are today. I'm not sure if a complete revert is the best next
>> step, or
Hi,
On 2019-01-14 18:03:24 -0500, Tom Lane wrote:
> Andres Freund writes:
> > On 2018-11-07 14:25:54 -0500, Tom Lane wrote:
> >> In short, it seems likely to me that large parts of this patch need to
> >> be pulled out, rewritten, and then put back in different places than
> >> they are today. I
Andres Freund writes:
> On 2019-01-14 18:03:24 -0500, Tom Lane wrote:
>> Do we want to revert entirely, or leave the "recheck_on_update" option
>> present but nonfunctional?
> I think it depends a bit on whether we want to revert in master or
> master and 11. If only master, I don't see much poin
On Mon, Jan 14, 2019 at 11:34 AM Tom Lane wrote:
> >> Hm. That would be annoying, but on reflection I think maybe we don't
> >> actually need to worry about emptiness of the array after all. The
> >> thing that we need to prove, at least for the implication case, is
> >> "truth of the ScalarArra
Hi,
On 2019-01-07 17:03:20 +0530, Amit Khandekar wrote:
> On Sat, 5 Jan 2019 at 02:09, Andres Freund wrote:
> > On 2019-01-03 13:40:42 -0500, Tom Lane wrote:
> > > I noticed that this patch has broken restores of existing dump files:
> > >
> > > psql:testbed.public.backup:82: ERROR: unrecognized
Greetings,
* Tom Lane (t...@sss.pgh.pa.us) wrote:
> Andres Freund writes:
> > On 2019-01-14 18:03:24 -0500, Tom Lane wrote:
> >> Do we want to revert entirely, or leave the "recheck_on_update" option
> >> present but nonfunctional?
>
> > I think it depends a bit on whether we want to revert in m
On 2019-01-14 18:46:18 -0500, Stephen Frost wrote:
> Greetings,
>
> * Tom Lane (t...@sss.pgh.pa.us) wrote:
> > Andres Freund writes:
> > > On 2019-01-14 18:03:24 -0500, Tom Lane wrote:
> > >> Do we want to revert entirely, or leave the "recheck_on_update" option
> > >> present but nonfunctional?
Stephen Frost writes:
> * Tom Lane (t...@sss.pgh.pa.us) wrote:
>> After a few minutes' more thought, I think that the most attractive
>> option is to leave v11 alone and do a full revert in HEAD. In this
>> way, if anyone's attached "recheck_on_update" options to their indexes,
>> it'll continue
Greetings,
* Andres Freund (and...@anarazel.de) wrote:
> On 2019-01-14 18:46:18 -0500, Stephen Frost wrote:
> > * Tom Lane (t...@sss.pgh.pa.us) wrote:
> > > Andres Freund writes:
> > > > On 2019-01-14 18:03:24 -0500, Tom Lane wrote:
> > > >> Do we want to revert entirely, or leave the "recheck_on
Hi,
On 2019-01-14 18:53:02 -0500, Tom Lane wrote:
> But I suspect just doing the revert is already going to be painful
> enough :-(
I assume you're not particularly interested in doing that? I'm more than
happy to leave this to others, but if nobody signals interest I'll give
it a go, because it'
Greetings,
* Tom Lane (t...@sss.pgh.pa.us) wrote:
> Stephen Frost writes:
> > * Tom Lane (t...@sss.pgh.pa.us) wrote:
> >> After a few minutes' more thought, I think that the most attractive
> >> option is to leave v11 alone and do a full revert in HEAD. In this
> >> way, if anyone's attached "re
Hi,
On 2019-01-14 18:55:18 -0500, Stephen Frost wrote:
> * Andres Freund (and...@anarazel.de) wrote:
> > > Or are you suggesting that pg_dump in v12+ would throw errors if it
> > > finds that set? Or that we'll dump it, but fail to allow it into a
> > > v12+ database? What if v12 sees "recheck_o
Andres Freund writes:
> On 2019-01-14 18:53:02 -0500, Tom Lane wrote:
>> But I suspect just doing the revert is already going to be painful
>> enough :-(
> I assume you're not particularly interested in doing that?
No, I'm willing to do it, and will do so tomorrow if there haven't
been objection
Greetings,
* Andres Freund (and...@anarazel.de) wrote:
> On 2019-01-14 18:55:18 -0500, Stephen Frost wrote:
> > * Andres Freund (and...@anarazel.de) wrote:
> > > > Or are you suggesting that pg_dump in v12+ would throw errors if it
> > > > finds that set? Or that we'll dump it, but fail to allow
Hi,
On 2019-01-14 19:04:07 -0500, Stephen Frost wrote:
> As that's the case, then I guess I'm thinking we really should make
> pg_upgrade complain if it finds it during the check phase. I really
> don't like having a case like this where the pg_upgrade will fail from
> something that we could hav
Greetings,
* Tom Lane (t...@sss.pgh.pa.us) wrote:
> Andres Freund writes:
> > On 2019-01-14 18:53:02 -0500, Tom Lane wrote:
> >> But I suspect just doing the revert is already going to be painful
> >> enough :-(
>
> > I assume you're not particularly interested in doing that?
>
> No, I'm willin
On Tue, 15 Jan 2019 at 09:48, Tom Lane wrote:
>
> David Rowley writes:
> > SELECT 1; I believe is a common query for some connection poolers as a
> > sort of ping to the database. In light of that, the performance drop
> > of 2 microseconds per query is not going to amount to very much in
> > to
Stephen Frost writes:
> * Tom Lane (t...@sss.pgh.pa.us) wrote:
>> What I'm not willing to do is write hacks for pg_upgrade or pg_dump
>> to mask cases where the option has been set on a v11 index. I judge
>> that it's not worth the trouble. If someone else disagrees, they
>> can do that work.
>
David Rowley writes:
> 1. I don't think having a table named "dual" makes a whole lot of
> sense for a table with a single row.
Well, I borrowed Oracle terminology there ;-)
> (Uppercasing these additions would also make them look less of an
> afterthought.)
Don't really care, can do.
> I als
On 2019/01/13 16:56, Michael Paquier wrote:
> On Fri, Jan 11, 2019 at 04:04:41PM +0900, Michael Paquier wrote:
>> Thanks. I can commit this version if there are no objections then.
>
> And done.
Thank you!
Regards,
Amit
From: Michael Paquier [mailto:mich...@paquier.xyz]
> One of the reasons why I have begun this thread is that since we have heard
> about the fsync issues on Linux, I think that there is room for giving our
> user base more control of their fate without relying on the Linux community
> decisions to
On Sun, Dec 23, 2018 at 9:33 AM Tomas Vondra
wrote:
> So, what is the expected speedup in a "good/average" case? Do we have
> some reasonable real-world workload mixing these cases that could be
> used as a realistic benchmark?
Not sure about a realistic mix, but I investigated the tradeoffs.
Fir
On Tue, 15 Jan 2019 at 13:17, Tom Lane wrote:
>
> David Rowley writes:
> > 1. I don't think having a table named "dual" makes a whole lot of
> > sense for a table with a single row.
>
> Well, I borrowed Oracle terminology there ;-)
yep, but ... go on... break the mould :)
--
David Rowley
From: Pavel Stehule [mailto:pavel.steh...@gmail.com]
> the cumulated lock statistics maybe doesn't help with debugging - but it
> is very good indicator of database (in production usage) health.
I think it will help both. But I don't think the sampling won't be as helpful
as the precise lock sta
Hi,
On 2019-01-14 17:55:44 -0300, Alvaro Herrera wrote:
> On 2019-Jan-14, Andres Freund wrote:
>
> > I think the whole pointer hiding game that we play is a really really
> > bad idea. We should just about *never* have typedefs that hide the fact
> > that something is a pointer. But it's hard to g
On 1/14/19 11:13 PM, Alvaro Herrera wrote:
> On 2019-Jan-14, Tomas Vondra wrote:
>
>> The one slightly annoying issue is that currently all the options are
>> formatted as text, including e.g. cpu_tuple_cost. That's because the GUC
>> options may have show hook, so I can't access the value dire
Fujita-san,
On 2019/01/11 20:04, Etsuro Fujita wrote:
> (2019/01/11 13:46), Amit Langote wrote:
>> On 2019/01/10 15:07, Etsuro Fujita wrote:
>>> (The name of the flag isn't
>>> good? If so, that would be my fault because I named that flag.)
>>
>> If it's really just to store the fact that the rel
On 2019/01/15 10:46, Amit Langote wrote:
> On 2019/01/11 20:04, Etsuro Fujita wrote:
>> One thing I intended in that commit was to set the flag to false for
>> partitioned tables contained in inheritance trees where the top parent is
>> a UNION ALL subquery, because we don't consider PWJ for those
On Tue, Nov 27, 2018 at 3:46 AM Petr Jelinek
wrote:
>
> Hi,
>
> On 26/11/2018 01:29, Masahiko Sawada wrote:
> > On Sun, Nov 25, 2018 at 12:27 AM Petr Jelinek
> > wrote:
> >>
> >> The more serious thing is:
> >>
> >>> + if (MyReplicationSlot)
> >>> + ReplicationSlotRelease();
> >>>
From: Tom Lane [mailto:t...@sss.pgh.pa.us]
> The problem here of course is that whoever invented target_session_attrs
> was unconcerned with following that precedent, so what we have is
> "target_session_attrs=(any | read-write)".
> Are we prepared to add some aliases in service of unifying these n
Hi,
I've started observing funny valgrind failures on Fedora 28, possibly
after upgrading from 3.14.0-1 to 3.14.0-7 a couple of days ago. This
time it does not seem like platform-specific issues, though - the
failures all look like this:
==20974== Conditional jump or move depends on uninitialised
Hi,
On 2019-01-15 03:07:10 +0100, Tomas Vondra wrote:
> I've started observing funny valgrind failures on Fedora 28, possibly
> after upgrading from 3.14.0-1 to 3.14.0-7 a couple of days ago. This
> time it does not seem like platform-specific issues, though - the
> failures all look like this:
A
On Sat, Jan 12, 2019 at 3:27 AM Robert Haas wrote:
>
> On Fri, Jan 11, 2019 at 1:14 AM Masahiko Sawada wrote:
> > Attached the updated patch. Please review it.
>
> I'm quite confused by this patch. It seems to me that the easiest way
> to implement this patch would be to (1) make lazy_space_allo
On Mon, Jan 14, 2019 at 07:31:07PM -0300, Alvaro Herrera wrote:
> On 2019-Jan-09, Justin Pryzby wrote:
>
> > -- Fails to error
> > postgres=# CREATE UNIQUE INDEX ON t(j) INCLUDE(i);
> >
> > -- Fail to enforce uniqueness across partitions due to failure to enforce
> > inclusion of partition key i
On Mon, Jan 14, 2019 at 07:41:18PM +0100, Dimitri Fontaine wrote:
> Thanks for the review here Tom, and for linking with the other
> discussion (Álvaro did that too, thanks!). I've been reviewing it
> too.
If you can look at the patch, reviews are welcome. There are quite a
couple of patterns I s
On 1/15/19 3:11 AM, Andres Freund wrote:
> Hi,
>
> On 2019-01-15 03:07:10 +0100, Tomas Vondra wrote:
>> I've started observing funny valgrind failures on Fedora 28, possibly
>> after upgrading from 3.14.0-1 to 3.14.0-7 a couple of days ago. This
>> time it does not seem like platform-specific issu
Fujita-san,
On 2019/01/11 21:50, Etsuro Fujita wrote:
>>> (2019/01/10 10:41), Amit Langote wrote:
That's a loaded meaning and abusing it to mean something else can be
challenged, but we can live with that if properly documented.
Speaking of
which:
/* used by pa
Hi,
On 2019-01-15 03:41:34 +0100, Tomas Vondra wrote:
> On 1/15/19 3:11 AM, Andres Freund wrote:
> > On 2019-01-15 03:07:10 +0100, Tomas Vondra wrote:
> >> I've started observing funny valgrind failures on Fedora 28, possibly
> >> after upgrading from 3.14.0-1 to 3.14.0-7 a couple of days ago. Thi
On Tue, Jan 15, 2019 at 02:00:57AM +, Tsunakawa, Takayuki wrote:
> But as some people said, I don't think this is the right way. I
> suppose what's leading to the current somewhat complicated situation
> is that there was no easy way for the client to know whether the
> server is the master.
Michael Paquier writes:
> On Tue, Jan 15, 2019 at 02:00:57AM +, Tsunakawa, Takayuki wrote:
>> The original desire should have been the ability to connect to a
>> primary or a standby. So, I think we should go back to the original
>> thinking (and not complicate the feature), and create a read
From: Tom Lane [mailto:t...@sss.pgh.pa.us]
> I'm really disappointed by the direction this thread is going in.
> The latest patches add an enormous amount of mechanism, and user-visible
> complexity, to do something that we learned was a bad idea decades ago.
> Putting a limit on the size of the sy
My nbtree patch [1] needs to call index_getprocinfo() with an
exclusive buffer lock held during a leaf page split. This has an
undetectable self-deadlock (LWLock self-deadlock) risk: a syscache
lookup against pg_proc might have a catcache miss, ending with an
index scan that needs to access the ver
Peter Geoghegan writes:
> My nbtree patch [1] needs to call index_getprocinfo() with an
> exclusive buffer lock held during a leaf page split.
I think you should stop right there and ask why. Surely that info
can be obtained before starting the operation? Quite aside from the
deadlock hazard, I
On Sun, Jan 13, 2019 at 10:17:32AM +0100, Peter Eisentraut wrote:
> I'm not sure what the coverage is in detail in this area. It seems we
> already have tests for not-specific-recovery-target, maybe not in this
> file, but most of the other tests rely on that, no?
[... Checking ...]
There are no
On Sat, 22 Dec 2018 at 07:10, Tom Lane wrote:
> BTW, with respect to this bit in 0001:
>
> @@ -1795,6 +1847,15 @@ nulltestsel(PlannerInfo *root, NullTestType
> nulltesttype, Node *arg,
> return (Selectivity) 0; /* keep compiler quiet */
> }
> }
> +else if (varda
Edmund Horner writes:
> On Sat, 22 Dec 2018 at 07:10, Tom Lane wrote:
>> I'm not entirely sure why you're bothering; surely nulltestsel is
>> unrelated to what this patch is about?
> I found that it made a difference with selectivity of range comparisons,
> because clauselist_selectivity tries t
On Mon, Jan 14, 2019 at 7:12 PM Tom Lane wrote:
> I think you should stop right there and ask why. Surely that info
> can be obtained before starting the operation?
*Thinks some more*
Uh, I'm already telling the same _bt_truncate() code path that it is
being called from a CREATE INDEX, allowing
current_logfiles is a meta data file, that stores the current log writing
file, and this file
presents in the data directory. This file doesn't follow the group access
mode set at
the initdb time, but it follows the log_file_mode permissions.
without group access permissions, backup with group acc
On 2019-Jan-14, Justin Pryzby wrote:
> On Mon, Jan 14, 2019 at 07:31:07PM -0300, Alvaro Herrera wrote:
> > Doh. Fix pushed. Commit 8224de4f42cc should have changed one
> > appearance of ii_NumIndexAttrs to ii_NumIndexKeyAttrs, but because of
> > the nature of concurrent development, nobody noti
I think, there's something better possible. Two partitioned relations won't
use partition-wise join, if their partition schemes do not match.
Partitioned relations with same partitioning scheme share PartitionScheme
pointer. PartitionScheme structure should get an extra counter, maintaining
a coun
On Tue, Jan 15, 2019 at 03:08:41PM +1100, Haribabu Kommi wrote:
> current_logfiles is a meta data file, that stores the current log writing
> file, and this file presents in the data directory. This file
> doesn't follow the group access mode set at the initdb time, but it
> follows the log_file_mo
On Tue, 15 Jan 2019 at 12:24, James Coleman wrote:
> While I add that though I wanted to get this updated version out to
> get feedback on the approach.
I had a look at this and there's a couple of things I see:
1. In:
+ if (IsA(clause, ScalarArrayOpExpr))
+ {
+ ScalarArrayOpExpr *saop = (Scala
On Sat, 12 Jan 2019 at 21:49, David Rowley wrote:
> > Can you say what you think is wrong with this way of making a copy of the
> > ECs?
>
> If you shallow copy with memcpy(new_ec, ec,
> sizeof(EquivalenceClass));, then fields such as ec_relids will just
> point to the same memory as the parent P
Hi Andrew,
This is my code to call the procedure with
SPI_connect_ext(SPI_OPT_NONATOMIC).
if (run_proc) {
appendStringInfo(&buf, "CALL \"%s\".run_maintenance_proc(p_analyze
:= %s, p_jobmon := %s);", partman_schema, analyze, jobmon);
expected_ret = SPI_OK_UTILITY;
function_
Thanks for the review and sorry it took me a while to reply.
On 2019/01/02 22:58, Peter Eisentraut wrote:
> On 26/11/2018 05:58, Amit Langote wrote:
>> On 2018/11/09 14:38, Amit Langote wrote:
>>> Rebased due to change in addRangeTableEntryForRelation's API.
>>
>> Rebased again due to changes in p
> On Mon, Jan 14, 2019 at 2:07 PM Amit Khandekar wrote:
>
> createPQExpBuffer() should be moved after the below statement, so that
> it does not leak memory
Thanks for noticing, fixed.
> So how about bumping up the archive version and doing these checks ?
Yeah, you're right, I've added this che
On Fri, Dec 28, 2018 at 11:43 AM Masahiko Sawada wrote:
>
> On Thu, Dec 20, 2018 at 3:38 PM Amit Kapila wrote:
> >
> > On Tue, Dec 18, 2018 at 1:29 PM Masahiko Sawada
> > wrote:
> > >
> > > Attached the updated patches. I scaled back the scope of this patch.
> > > The patch now includes only fe
On Tue, Dec 11, 2018 at 1:13 PM Andres Freund wrote:
> Hi,
>
> On 2018-11-26 17:55:57 -0800, Andres Freund wrote:
> Further tasks I'm not yet planning to tackle, that I'd welcome help on:
> - pg_upgrade testing
>
I did the pg_upgrade testing from older version with some tables and views
exists,
Hi,
On 2019-01-15 18:02:38 +1100, Haribabu Kommi wrote:
> On Tue, Dec 11, 2018 at 1:13 PM Andres Freund wrote:
>
> > Hi,
> >
> > On 2018-11-26 17:55:57 -0800, Andres Freund wrote:
> > Further tasks I'm not yet planning to tackle, that I'd welcome help on:
> > - pg_upgrade testing
> >
>
> I did
On Sun, Jan 13, 2019 at 03:31:23PM +0100, Pavel Stehule wrote:
> ne 13. 1. 2019 v 10:43 odesílatel Peter Eisentraut <
> peter.eisentr...@2ndquadrant.com> napsal:
>> See here:
>>
>> https://www.postgresql.org/message-id/b5c27634-1d44-feba-7494-ce5a31f91...@2ndquadrant.com
>
> I understand - it is l
1 - 100 of 101 matches
Mail list logo