First, thanks for taking the time to write this. Its very helpful.
Additional thoughts inline.
On Mon, Sep 24, 2018 at 2:12 AM Michael Paquier wrote:
> On Fri, Sep 21, 2018 at 12:35:46PM +0200, Chris Travers wrote:
> > I understand how lock levels don't fit a simple hierarchy but at least
> > w
Hi!
I want to review this patch set. Though I understand that it probably will be
quite long process.
I like the idea that with this patch set universally all postgres instances are
bound into single distributed DB, even if they never heard about each other
before :) This is just amazing. Or d
On Mon, Sep 24, 2018 at 8:04 AM Haribabu Kommi wrote:
> On Mon, Sep 24, 2018 at 5:02 AM Alexander Korotkov
> wrote:
>>
>> On Fri, Aug 24, 2018 at 5:50 AM Andres Freund wrote:
>> > I've pushed a current version of that to my git tree to the
>> > pluggable-storage branch. It's not really a versio
I did some more reading.
On Mon, Sep 24, 2018 at 10:15 AM Chris Travers
wrote:
> First, thanks for taking the time to write this. Its very helpful.
> Additional thoughts inline.
>
> On Mon, Sep 24, 2018 at 2:12 AM Michael Paquier
> wrote:
>
>>
>> There could be value in refactoring things so a
On Thu, Sep 20, 2018 at 12:00:18PM +0200, Marco Slot wrote:
> We're seeing a segmentation fault when creating a partition of a
> partitioned table with a primary key when there is a sql_drop trigger on
> Postgres 11beta4.
Thanks for reporting ; I reproduced easily so added to open items list, sinc
On 11/09/2018 17:10, Peter Eisentraut wrote:
> On 07/09/2018 17:59, Maksim Milyutin wrote:
>> those directories was that user). The error message "could not set
>> permissions on directory ..." disoriented that user. The need to change
>> the owner of those directories came after careful reading
On 30/07/2018 13:51, Jeff Janes wrote:
> Any thoughts on how to proceed here? It seems there is more work to do
> to cover all the issues with dumping and restoring tables with many
> columns. Since the original report was in the context of pg_upgrade, we
> should surely address a
Very minor revision
On Mon, Sep 24, 2018 at 11:45 AM Chris Travers
wrote:
>
> Ok so having looked into this a bit more
>
> It looks like to be strictly conforming, you can't just use a series of
> flags because neither C 89 or 99 guarantee that sig_atomic_t is read/write
> round-trip safe in
On 09/21/2018 01:51 PM, Joe Conway wrote:
> On 09/19/2018 11:18 AM, Joe Conway wrote:
>> On 09/19/2018 10:54 AM, Tom Lane wrote:
>>> So maybe what we really need is a table of operators not functions.
>>
>> Good idea -- I will take a look at that.
>>
>>> However, I don't object to documenting any
>
> The quotas or object limits, resource limits are pretty useful and
> necessary, but I don't see these like new type of objects, it is much more
> some property of current objects. Because we have one syntax for this
> purpose I prefer it. Because is not good to have two syntaxes for similar
> p
Joe Conway writes:
> Having seen none, committed/pushed. This did not seem worth
> back-patching, so I only pushed to master.
I don't see anything on gitmaster?
regards, tom lane
On 29/04/2018 20:18, Joe Wildish wrote:
> On 28 Mar 2018, at 16:13, David Fetter wrote:
>>
>> Sorry to bother you again, but this now doesn't compile atop master.
>
> Attached is a rebased patch for the prototype.
I took a look at this.
This has been lying around for a few months, so it will ne
On 09/24/2018 10:01 AM, Tom Lane wrote:
> Joe Conway writes:
>> Having seen none, committed/pushed. This did not seem worth
>> back-patching, so I only pushed to master.
>
> I don't see anything on gitmaster?
Hmm, yes, interesting -- I must of messed up my local git repo somehow.
Will try again.
On 09/24/2018 10:09 AM, Joe Conway wrote:
> On 09/24/2018 10:01 AM, Tom Lane wrote:
>> Joe Conway writes:
>>> Having seen none, committed/pushed. This did not seem worth
>>> back-patching, so I only pushed to master.
>>
>> I don't see anything on gitmaster?
>
> Hmm, yes, interesting -- I must of
Keiichi Hirobe writes:
> Attached is a patch that fixes a bug
> for miscounting total number of curly braces in output string in array_out.
Wow, good catch!
Testing this, I found there's a second way in which the space calculation
is off: it always allocated one more byte than required, as a res
On Mon, Sep 24, 2018 at 12:19:44PM +1000, Haribabu Kommi wrote:
> Attached new rebased version of the patch that enhances the
> pg_stat_statements_reset()
> function. This needs to be applied on top of the patch that is posted in
> [1].
+CREATE ROLE stats_regress_user1;
+CREATE ROLE stats_regress_
On Mon, Sep 24, 2018 at 11:08:14AM +1000, Haribabu Kommi wrote:
> In commit 25fff40798 the execute permission of pg_stat_statements_reset()
> is provided to pg_read_all_stats role in [1].
>
> The execute permissions grant to pg_read_all_stats concern is raised in [2]
> during the discussion of sup
[Back from a week AFK ...]
On 2018-Sep-18, amul sul wrote:
> On Mon, Sep 17, 2018 at 9:06 PM amul sul wrote:
> >
> > Nice catch Rajkumar.
> Here is the complete patch proposes the aforesaid fix with regression test.
Looks closely related to the one that was fixed in 1f8a3327a9db, but of
course
On 2018-Sep-20, Marco Slot wrote:
> We're seeing a segmentation fault when creating a partition of a
> partitioned table with a primary key when there is a sql_drop trigger on
> Postgres 11beta4.
>
> We discovered it because the Citus extension creates a sql_drop trigger,
> but it's otherwise unr
Michael Paquier writes:
> This should be back-patched. Any opinions about bumping up this
> extension version in back-branches like what has been done in 53b79ab4?
Yes, you need to bump the extension version to change anything in the
extension's script file.
For v10 and up, the method used in 5
Thomas Munro writes:
> On Mon, Sep 24, 2018 at 9:32 AM Tom Lane wrote:
>> Mark G writes:
>>> While looking at some of the recent churn in DSA I noticed that
>>> dsa_size_class_map should probably be declared const.
>> +1 ... also, given the contents of the array, "char" seems like
>> rather a m
It's been a bunch of years since I worked with ICU, so anything I say below
may have changed in their code or be subject to mental bit rot.
On Sun, Sep 23, 2018 at 2:48 PM Thomas Munro
wrote:
> Considering that to handle this we'd need to figure out
> how link libicu.so.55, libicu.so.56, ... etc
On Thursday, March 1, 2018 4:25:16 AM EDT Andres Freund wrote:
> A CF entry for this patch has been created yesterday, without any
> changes having happend since the above status update. Given that
> there's been no progress for multiple commitfests, and this is the
> last CF, this doesn't seem ap
On 2018-09-24 11:45:10 +0200, Chris Travers wrote:
> I did some more reading.
>
> On Mon, Sep 24, 2018 at 10:15 AM Chris Travers
> wrote:
>
> > First, thanks for taking the time to write this. Its very helpful.
> > Additional thoughts inline.
> >
> > On Mon, Sep 24, 2018 at 2:12 AM Michael Paqu
Michael Paquier writes:
> On Wed, Sep 12, 2018 at 01:40:01PM -0400, Tom Lane wrote:
>> Rebase attached --- no substantive changes.
> - if (handleDLL == NULL)
> - ereport(FATAL,
> - (errmsg_internal("could not load netmsg.dll: error
> -code %lu", GetLa
On Wed, Sep 19, 2018 at 11:23 AM Peter Geoghegan wrote:
> 3 modes
> ---
>
> My new approach is to teach _bt_findsplitloc() 3 distinct modes of
> operation: Regular/default mode, many duplicates mode, and single
> value mode.
I think that I'll have to add a fourth mode, since I came up with
an
Noah Misch writes:
> On Tue, May 01, 2018 at 11:31:50AM -0400, Tom Lane wrote:
>> Well, at this point the only thing that's entirely clear is that none
>> of the ideas I had work. I think we are going to be forced to pursue
>> Noah's idea of doing an end-to-end retry. Somebody else will need to
Hi,
On 2018-09-19 20:39:22 -0700, Andres Freund wrote:
> On 2018-09-19 23:26:52 -0400, Tom Lane wrote:
> > That's going in the right direction. Personally I'd make the last line
> > more like
> >
> > Times: generation 0.680 ms, inlining 7.591 ms, optimization 20.522 ms,
> > emission 14.607 m
On 2018-09-24 18:25:46 +, Nasby, Jim wrote:
>
> > On Sep 21, 2018, at 12:43 PM, Andres Freund wrote:
> >
> >> But as far i can see it is possible have aggressive non-wraparound vacuum.
> >> One important difference - regular and aggressive regular can be canceled
> >> by backend,.wraparoun
Hi,
Whilst playing around with auto_explain and JIT today, I realized that
auto_explain currently doesn't output JIT information, which is rather
unfortunate when analyzing a larger set of queries in a semi-automated
manner.
Attached a trivial patch that fixes the issue and adds JIT information t
Tahir,
I've extended an invitation through GCI for you to become a mentor for the
PostgreSQL
organization, and have sent an additional email containing information on
how to get started
with contributing to this year's contest. We have one month left to get
over a hundred tasks
uploaded, so that i
On Tue, Sep 25, 2018 at 4:17 AM Tom Lane wrote:
> Thomas Munro writes:
> > On Mon, Sep 24, 2018 at 9:32 AM Tom Lane wrote:
> >> Mark G writes:
> >>> While looking at some of the recent churn in DSA I noticed that
> >>> dsa_size_class_map should probably be declared const.
>
> >> +1 ... also, gi
Hi,
(CCing -hackers)
On 2018-09-24 12:22:28 -0700, legrand legrand wrote:
> I have found that explain on tables with many (hundreds) columns
> are slow compare to nominal executions.
Yea, colname_is_unique() (called via make_colname_unique()) is
essentially O(#total_columns) and rougly called on
Thanks Sarah,
I have received the emails and invitation. Now, I am reviewing all stuff
and will respond with my ideas within a day.
Regards
Tahir
On Tue, Sep 25, 2018 at 12:34 AM Sarah Conway Schnurr <
xenophene...@gmail.com> wrote:
> Tahir,
>
> I've extended an invitation through GCI for you t
> On Sep 24, 2018, at 1:29 PM, Andres Freund wrote:
>
> I'm very doubtful this is an improvement. Especially with the upcoming
> pluggable storage work making vacuumlazy.c heap specific, while vacuum.c
> stays generic. The concept of something like
> PROC_VACUUM_FOR_WRAPAROUND, should imo not b
On 09/20/2018 09:04 AM, Tomas Vondra wrote:
On 09/19/2018 10:35 PM, Andrew Dunstan wrote:
On 09/18/2018 03:36 PM, Andrew Dunstan wrote:
Tomas Vondra has pointed out to me that there's an issue with
triggers not getting expanded tuples for columns with fast defaults.
Here is an exampl
On Tue, Sep 25, 2018 at 4:26 AM Douglas Doole wrote:>
> On Sun, Sep 23, 2018 at 2:48 PM Thomas Munro
> wrote:
>> Admittedly that creates a whole can
>> of worms for initdb-time catalog creation, package maintainers' jobs,
>> how long old versions have to be supported and how you upgraded
>> data
Hi,
On 2018-09-24 11:34:38 -0700, Lukas Fittl wrote:
> Hi,
>
> Whilst playing around with auto_explain and JIT today, I realized that
> auto_explain currently doesn't output JIT information, which is rather
> unfortunate when analyzing a larger set of queries in a semi-automated
> manner.
>
> At
Hi
> An autovacuum can't be just aggressive; it's either anti-wraparound or normal.
But autovacuum _can_ be aggressive and not anti-wraparound.
I build current master and can see 3 different line types:
2018-09-24 23:47:31.500 MSK 27939 @ from [vxid:4/272032 txid:0] [] LOG:
automatic aggressive
Greetings Don!
* Don Seiler (d...@seiler.us) wrote:
> On Tue, Aug 7, 2018 at 12:32 PM Tom Lane wrote:
> > Don Seiler writes:
> >
> > > 1. We want to make a generic, central ascii-lobotomizing function similar
> > > to check_application_name that we can re-use there and for other checks
> > (eg
>
On 2018-Sep-24, Sergei Kornilov wrote:
> Hi
>
> > An autovacuum can't be just aggressive; it's either anti-wraparound or
> > normal.
> But autovacuum _can_ be aggressive and not anti-wraparound.
> I build current master and can see 3 different line types:
> 2018-09-24 23:47:31.500 MSK 27939 @ fr
On Mon, Sep 24, 2018 at 1:47 PM Thomas Munro
wrote:
> Personally I'm not planning to work on multi-version installation any
> time soon, I was just scoping out some basic facts about all this. I
> think the primary problem that affects most of our users is the
> shifting-under-your-feet problem,
Arthur Zakirov writes:
> Ah, I see. I attached new version made with --no-renames. Will wait for
> what cfbot will say.
I reviewed and pushed this.
As a cross-check on the patch, I cloned the Snowball github repo
and built the derived files in it. I noticed that they'd incorporated
several new
On Mon, Sep 24, 2018 at 1:48 PM, Andres Freund wrote:
>
> Thanks for noticing - pushed!
>
Thanks!
Best,
Lukas
--
Lukas Fittl
David Rowley said:
> I believe that we should be delaying the PlannerInfo's
> total_table_pages calculation until after constraint exclusion and
> partition pruning have taken place. Doing this calculation before we
> determine which relations we don't need to scan can lead to
> incorrectly applyin
Hi Peter,
> On 24 Sep 2018, at 15:06, Peter Eisentraut
> wrote:
>
> On 29/04/2018 20:18, Joe Wildish wrote:
>>
>> Attached is a rebased patch for the prototype.
>
> I took a look at this.
Thank you for reviewing.
> This has been lying around for a few months, so it will need to be
> rebased
Thanks for the feedback.
On 2018-09-22, Amit Kapila wrote ...
> ... Why can't we just extend the current Note where it is currently
...
Because information about how the protocol works belongs in the protocol
documentation not in the documentation for one implementation of the
protocol.
Th
On Mon, Sep 24, 2018 at 10:06:40AM -0700, Andres Freund wrote:
> This doesn't seem to solve an actual problem, why are we discussing
> changing this? What'd be measurably improved, worth the cost of making
> backpatching more painful?
My point was just to reduce the number of variables used and ea
On 2018-09-25 08:57:25 +0900, Michael Paquier wrote:
> On Mon, Sep 24, 2018 at 10:06:40AM -0700, Andres Freund wrote:
> > This doesn't seem to solve an actual problem, why are we discussing
> > changing this? What'd be measurably improved, worth the cost of making
> > backpatching more painful?
>
> "Joe" == Joe Wildish writes:
Joe> Agreed. My assumption was that we would record in the data
Joe> dictionary the behaviour (or “polarity") of each aggregate
Joe> function with respect to the various operators. Column in
Joe> pg_aggregate? I don’t know how we’d record it exactly.
I have
On Mon, Sep 24, 2018 at 12:02:35PM -0400, Tom Lane wrote:
> For v10 and up, the method used in 53b79ab4 is overcomplicated: you only
> need to add a delta script not a new base script. (If you had to
> back-patch before v10, it might be best to add a new base script in all
> the branches just to k
On Mon, Sep 24, 2018 at 05:23:56PM -0700, Andres Freund wrote:
> On 2018-09-25 08:57:25 +0900, Michael Paquier wrote:
>> Anyway, putting the back-patching pain aside, and just for my own
>> knowledge... Andres, would it be fine to just use one sig_atomic_t
>> field which can be set from different
Michael Paquier writes:
> And then within separate signal handlers things like:
> void
> StatementCancelHandler(SIGNAL_ARGS)
> {
> [...]
> signalPendingFlags |= PENDING_INTERRUPT | PENDING_CANCEL_QUERY;
> [...]
> }
AFAICS this still wouldn't work. The machine code is still going to
l
On Mon, Sep 24, 2018 at 09:03:49PM -0400, Tom Lane wrote:
> You could only fix that by blocking all signal handling during the
> handler, which would be expensive and rather pointless.
>
> I do not think that it's readily possible to improve on the current
> situation with one sig_atomic_t per fla
Michael Paquier writes:
> At the same time, all the pending flags in miscadmin.h could be switched
> to sig_atomic_t if we were to be correct, no? The counters could be
> higher than 256 so that's not really possible.
Yeah, in principle any global variable touched by a signal handler should
be
On Fri, Sep 21, 2018 at 05:44:02PM +0200, Gilles Darold wrote:
> Attached is a patch that fixes a bug in pg_dump since 10.0 and
> reproducible in master. When using option --no-publication : ALTER
> PUBLICATION orders are still present in the dump.
Thanks for the report, the patch, and the test ca
Hello hackers,
Some people like to use DNS SRV records to advertise LDAP servers on
their network. Microsoft Active Directory is usually (always?) set up
that way. Here is a patch to allow our LDAP auth module to support
that kind of discovery. It copies the convention of the OpenLDAP
command l
From: Michael Paquier [mailto:mich...@paquier.xyz]
> Okay, I have pushed the patch with all your suggestions included.
Thanks so much!
Regards
Takayuki Tsunakawa
On Tue, Sep 25, 2018 at 2:09 PM Thomas Munro
wrote:
> 2. Define a new zone for testing, by adding the following to the end
> 3. Create that zone file in /usr/local/etc/namedb/master/my.test.domain:
Oops, I changed my testing domain name in the middle of my experiment,
but pasted the older versi
On Mon, Sep 24, 2018 at 09:38:11PM -0400, Tom Lane wrote:
> Yeah, in principle any global variable touched by a signal handler should
> be sig_atomic_t. I don't know of any modern platform where using "bool"
> is unsafe, but per the C standard it could be. The case that would be
> worrisome is if
Michael Paquier writes:
> Let's change it then. ClientConnectionLost needs also to be changed as
> miscadmin.h tells that it could be used in a signal handler. What do you
> think about the attached?
Looks reasonable to me (I've not tested though).
I wonder why ClientConnectionLost isn't PGDLLI
On Mon, Sep 24, 2018 at 10:39:40PM -0400, Tom Lane wrote:
> I wonder why ClientConnectionLost isn't PGDLLIMPORT like the rest.
Same question here. As that's kind of a separate discussion, I left it
out. Now I don't mind changing that at the same time as that's
harmless, and as that's only a patc
On Tue, Sep 25, 2018 at 7:46 AM Thomas Munro
wrote:
> On Tue, Sep 25, 2018 at 4:17 AM Tom Lane wrote:
> > Thomas Munro writes:
> > > On Mon, Sep 24, 2018 at 9:32 AM Tom Lane wrote:
> > >> Mark G writes:
> > >>> While looking at some of the recent churn in DSA I noticed that
> > >>> dsa_size_cl
On Tue, Sep 25, 2018 at 10:58 AM Michael Paquier
wrote:
> On Mon, Sep 24, 2018 at 12:02:35PM -0400, Tom Lane wrote:
> > For v10 and up, the method used in 53b79ab4 is overcomplicated: you only
> > need to add a delta script not a new base script. (If you had to
> > back-patch before v10, it migh
On Tue, Sep 25, 2018 at 1:39 AM Michael Paquier wrote:
> On Mon, Sep 24, 2018 at 12:19:44PM +1000, Haribabu Kommi wrote:
> > Attached new rebased version of the patch that enhances the
> > pg_stat_statements_reset()
> > function. This needs to be applied on top of the patch that is posted in
> >
Hello.
At Mon, 17 Sep 2018 22:13:40 +0900, Michael Paquier wrote
in <20180917131340.ge31...@paquier.xyz>
> Hi all,
>
> On a rather freshly-updated Debian SID server, I am able to see failures
> for the SSL TAP tests:
> 2018-09-17 22:00:27.389 JST [13072] LOG: database system is shut down
> 201
So I've tried to rough out a decision tree for the various options on
how this might be implemented (discarding the "use precedence hacks"
option). Opinions? Additions?
(formatted for emacs outline-mode)
* 1. use lexical lookahead
+: relatively straightforward parser changes
+: no new reserv
On 2018-09-25 11:50:47 +0900, Michael Paquier wrote:
> PGDLLIMPORT changes don't get back-patched as well...
We've been more aggressive with that lately, and I think that's good. It
just is a unnecessarily portability barrier for extension to be
judicious in applying that.
- Andres
Sawada-san,
On Mon, Sep 24, 2018 at 08:28:45AM +0900, Michael Paquier wrote:
> Wouldn't it be better to incorporate the new test as part of
> 004_restart.pl? This way, we avoid initializing a full instance, which
> is always a good thing as that's very costly. The top of this file also
> mention
Well, macOS 10.14 (Mojave) is out, so I installed it on a spare machine,
and naturally the first thing I tried was to build PG with it. Our
core code seems fine, but:
* --with-perl fails in configure, complaining that it can't find perl.h.
* --with-tcl fails in configure, complaining that it can
On Tue, Sep 25, 2018 at 01:49:09PM +1000, Haribabu Kommi wrote:
> Thanks for the review.
> Fixed in the attached patch as per your suggestion.
Hmm. I see a problem with the tests and the stability of what
pg_stat_statements_reset() can return. Normally installcheck is
disabled in contrib/pg_stat
On Tue, Sep 25, 2018 at 12:48:57PM +0900, Kyotaro HORIGUCHI wrote:
> Do you mean that cert/key files are generated on-the-fly while
> running 'make check'? It sounds reasonable as long as just
> replaceing existing files with those with longer (2048bits?) keys
> doesn't work for all supported plat
I was about to suggest creating a single shared snapshot instead of having
multiple backends compute what is essentially the same snapshot. Luckily,
before posting, I discovered Avoiding repeated snapshot computation
https://www.postgresql.org/message-id/caboikdmsj4osxta7xbv2quhkyuo_4105fjf4n+u
On Tue, Sep 25, 2018 at 11:00 AM, Daniel Wood wrote:
> I was about to suggest creating a single shared snapshot instead of having
> multiple backends compute what is essentially the same snapshot. Luckily,
> before posting, I discovered Avoiding repeated snapshot computation from
> Pavan and POC:
On Sat, Sep 22, 2018 at 4:52 PM Amit Kapila wrote:
> On Sat, Sep 22, 2018 at 2:28 AM Andres Freund wrote:
> > On 2018-09-22 08:54:57 +1200, Thomas Munro wrote:
> > > On Fri, Sep 21, 2018 at 4:43 PM Tom Lane wrote:
> > > > Unless it looks practical to support this behavior in the Windows
> > > >
On Tue, Sep 25, 2018 at 3:03 AM Tom Lane wrote:
> Michael Paquier writes:
> > And then within separate signal handlers things like:
> > void
> > StatementCancelHandler(SIGNAL_ARGS)
> > {
> > [...]
> > signalPendingFlags |= PENDING_INTERRUPT | PENDING_CANCEL_QUERY;
> > [...]
> > }
>
>
76 matches
Mail list logo
