On 26 August 2017 at 05:40, Mark Kirkwood wrote:
> On 26/08/17 12:18, Simon Riggs wrote:
>
>> On 25 August 2017 at 20:53, Tom Lane wrote:
>>>
>>> Greg Stark writes:
I think this is a particularly old piece of code and we're lucky the
default heuristics have served well for all thi
On Mon, Aug 28, 2017 at 9:06 PM, Michael Paquier
wrote:
> On Tue, Jun 13, 2017 at 4:21 AM, Andres Freund wrote:
>> On 2017-06-12 15:12:23 -0400, Robert Haas wrote:
>>> Commit 4b4b680c3d6d8485155d4d4bf0a92d3a874b7a65 (Make backend local
>>> tracking of buffer pins memory efficient., vintage 2014)
On Tue, Aug 29, 2017 at 10:16 AM, Robert Haas wrote:
> On Mon, Aug 28, 2017 at 5:26 AM, Kyotaro HORIGUCHI
> wrote:
>> Currently the message shows the '%d skipped-frozen' message but
>> it is insufficient to verify the true effect. This is a patch to
>> show mode as 'aggressive' or 'normal' in the
Cache the SnapshotData for reuse:
===
In one of our perf analysis using perf tool it showed GetSnapshotData
takes a very high percentage of CPU cycles on readonly workload when
there is very high number of concurrent connections >= 64.
Machine : cthulhu 8 node machine.
Hi hackers!
> 8 авг. 2017 г., в 11:27, Tom Lane написал(а):
>
> My point is not to claim that we mustn't put a hook there. It's that what
> such a hook could safely do is tightly constrained, and you've not offered
> clear evidence that there's something useful to be done within those
> constra
On Wed, Aug 23, 2017 at 9:45 AM, Dilip Kumar wrote:
>>
>> ...
>> if (tbm->nentries <= tbm->maxentries / 2)
>> {
>> /*
>> * We have made enough room.
>> ...
>> I think we could try higher fill factor, say, 0.9. tbm_lossify basically
>> just
On Sat, Aug 26, 2017 at 8:00 AM, Robert Haas wrote:
> What happens if you say VACUUM partitioned_table (a), some_partition (b)?
Using v9, if you do that:
=# CREATE TABLE parent (id int) PARTITION BY RANGE (id);
CREATE TABLE
=# CREATE TABLE child_1_10001 partition of parent for values from (1)
to
On Mon, Aug 28, 2017 at 5:50 PM, Tom Lane wrote:
> I think that's probably dead code given that ExecutorRun short-circuits
> everything for NoMovementScanDirection. There is some use of
> NoMovementScanDirection for indexscans, to denote an unordered index,
> but likely that could be got rid of t
On Mon, Aug 28, 2017 at 2:26 AM, Kyotaro HORIGUCHI <
horiguchi.kyot...@lab.ntt.co.jp> wrote:
> https://www.postgresql.org/docs/devel/static/runtime-config-client.html
>
> >
> V
> ACUUM performs an aggressive scan
>
Maybe this should gets its own thread/patch but I'll tack this on here
since it
On Mon, Aug 28, 2017 at 10:17 PM, Tom Lane wrote:
> In the meantime, I think what we should do is commit the bug fix more or
> less as I have it, and then work on Amit's concern about losing parallel
> efficiency by separating the resetting of shared parallel-scan state
> into a new plan tree trav
On Mon, Aug 28, 2017 at 10:44 PM, yangjie wrote:
> When the number of partitions and the data are more, adding new partitions,
> there will be some efficiency problems.
> I don't know how the solution you're talking about is how to implement a
> hash partition?
I am having difficulty understandin
Michael Paquier writes:
> I don't like breaking the abstraction of pg_log() with the existing
> flags with some kind of pg_debug() layer. The set of APIs present now
> in pg_rewind for logging is nice to have, and I think that those debug
> messages should be translated. So what about the attached
On Mon, Aug 28, 2017 at 11:24 PM, Robert Haas wrote:
> On Mon, Aug 28, 2017 at 10:16 AM, Alvaro Herrera
> wrote:
>>> I am fine with however you want to handle it, but it seems odd to me
>>> that we don't have a way of embedding INT64_FORMAT in a translatable
>>> string. Surely that's going to be
Robert Haas writes:
> On Mon, Aug 28, 2017 at 6:35 PM, Tom Lane wrote:
>> Maybe parallel_aware should have more than two values, depending
>> on whether the result of the node is context-dependent or not.
> It seems likely the whole concept of parallel_aware is only only a
> zero-order approxima
On Mon, Aug 28, 2017 at 5:24 AM, Kyotaro HORIGUCHI
wrote:
> This patch have had interferences from several commits after the
> last submission. I amended this patch to follow them (up to
> f97c55c), removed an unnecessary branch and edited some comments.
I think the core problem for this patch is
On Mon, Aug 28, 2017 at 3:05 PM, Michael Paquier
wrote:
> Attached are two patches:
> 1) 0001 refactors the code around pqAddTuple to be able to handle
> error messages and assign them in PQsetvalue particularly.
> 2) 0002 adds sanity checks in pqAddTuple for overflows, maximizing the
> size of wh
On Mon, Aug 28, 2017 at 4:59 PM, Fabien COELHO wrote:
>
> Hello Masahiko-san,
>
> Patch applies cleanly, compiles, works for me.
Thank you for reviewing!
>
>>> At least: "Required to invoke" -> "Require to invoke".
>>
>>
>> Fixed.
>
>
> I meant the added one about -I, not the pre-existing one ab
On 29 August 2017 at 05:15, Tom Lane wrote:
> Peter Eisentraut writes:
> > On 8/28/17 15:11, Tom Lane wrote:
> >> ... but it seems like you're giving up a lot of the possible uses if
> >> you don't make it apply uniformly. I admit I'm not sure how we'd handle
> >> the initial creation of a conn
On Mon, Aug 28, 2017 at 5:26 AM, Kyotaro HORIGUCHI
wrote:
> Currently the message shows the '%d skipped-frozen' message but
> it is insufficient to verify the true effect. This is a patch to
> show mode as 'aggressive' or 'normal' in the closing message of
> vacuum. %d frozen-skipped when 'aggress
On Mon, Aug 28, 2017 at 12:23 PM, Antonin Houska wrote:
> Now that dynamic shared memory hash table has been committed
> (8c0d7bafad36434cb08ac2c78e69ae72c194ca20) I wonder if it's still a big deal
> to remove restrictions like this in (e.g. heap_update()):
>
> /*
> * Forbid this
On Mon, Aug 28, 2017 at 6:35 PM, Tom Lane wrote:
> but probably we should think of a more arm's-length way to do it.
> Maybe parallel_aware should have more than two values, depending
> on whether the result of the node is context-dependent or not.
My original intent for the parallel_aware flag w
On Thu, Jul 6, 2017 at 07:29:07AM -0700, Jim Finnerty wrote:
> re: "The problem is if you want to delete from such a page. Then you need to
> update the tuple's xmax and stick the new xid epoch somewhere."
I am coming to this very late, but wouldn't such a row be marked using
our frozen-commited
Robert Haas writes:
> On Mon, Aug 28, 2017 at 3:00 PM, Tom Lane wrote:
>> Count the "that"s (you're failing to notice the next line).
> OK, true. But "Academic literature" -> "The academic literature" is
> just second-guessing, I think.
No, it was more to avoid reflowing the paragraph (or leav
On Sat, Aug 26, 2017 at 8:00 AM, Robert Haas wrote:
> On Thu, Aug 24, 2017 at 12:59 AM, Michael Paquier
> wrote:
>> Robert, Amit and other folks working on extending the existing
>> partitioning facility would be in better position to answer that, but
>> I would think that we should have somethin
On Mon, Aug 28, 2017 at 3:00 PM, Tom Lane wrote:
>> That sentence isn't wrong as written.
>
> Count the "that"s (you're failing to notice the next line).
OK, true. But "Academic literature" -> "The academic literature" is
just second-guessing, I think.
>> I don't really understand the part abou
Peter Eisentraut writes:
> On 8/28/17 15:11, Tom Lane wrote:
>> ... but it seems like you're giving up a lot of the possible uses if
>> you don't make it apply uniformly. I admit I'm not sure how we'd handle
>> the initial creation of a connection object with a custom malloc. The
>> obvious solu
On 8/28/17 15:11, Tom Lane wrote:
> ... but it seems like you're giving up a lot of the possible uses if
> you don't make it apply uniformly. I admit I'm not sure how we'd handle
> the initial creation of a connection object with a custom malloc. The
> obvious solution of requiring the functions
On Mon, Aug 28, 2017 at 03:11:26PM -0400, Tom Lane wrote:
> Aaron Patterson writes:
> > I would like to be able to configure libpq with custom malloc functions.
>
> I can see the potential value of this ...
>
> > This patch doesn't replace all malloc calls to the configured ones, just
> > the ma
On Mon, Aug 28, 2017 at 6:38 AM, Amit Langote
wrote:
> I am worried about the open relcache reference in PartitionDispatch when
> we start using it in the planner. Whereas there is a ExecEndModifyTable()
> as a suitable place to close that reference, there doesn't seem to exist
> one within the p
Aaron Patterson writes:
> I would like to be able to configure libpq with custom malloc functions.
I can see the potential value of this ...
> This patch doesn't replace all malloc calls to the configured ones, just
> the mallocs related to creating result objects (which is what I'm
> concerned
Robert Haas writes:
> On Mon, Aug 28, 2017 at 10:47 AM, Tom Lane wrote:
> - fuller description. Academic literature on parallel query suggests that
> + fuller description. The academic literature on parallel query suggests
> That sentence isn't wrong as written.
Count the "that"s (you're fail
On Mon, Aug 28, 2017 at 10:47 AM, Tom Lane wrote:
> That seems like an unacceptably fragile assumption. Even if it happens to
> be true today, we would need to fix it sooner or later. (And I kinda
> suspect it's possible to break it today, anyway. Treating PARAM_EXEC
> Params as parallel-restri
Hello!
I would like to be able to configure libpq with custom malloc functions.
The reason is that we have a Ruby wrapper that exposes libpq in Ruby.
The problem is that Ruby's GC doesn't know how much memory has been
allocated by libpq, so no pressure is applied to the GC when it should
be. Ruby
Now that dynamic shared memory hash table has been committed
(8c0d7bafad36434cb08ac2c78e69ae72c194ca20) I wonder if it's still a big deal
to remove restrictions like this in (e.g. heap_update()):
/*
* Forbid this during a parallel operation, lest it allocate a combocid.
*
On 08/24/2017 03:08 PM, Maksim Milyutin wrote:
[...]
>
> AFAICS you want to introduce two additional per-node variables:
> - auto_explain_log_estimate_ratio that denotes minimum ratio (>= 1) between
> real value and planned one. I would add 'min' prefix before 'ratio'.
> - auto_explain_log_estim
Amit Kapila writes:
> On Mon, Aug 28, 2017 at 6:34 PM, Tom Lane wrote:
>> If what you're complaining about is that I put back the "if
>> (outerPlan->chgParam == NULL)" test to allow postponement of the
>> recursive ExecReScan call, I'm afraid that it's mere wishful
>> thinking that omitting that
> Thanks for reporting it!
>
My pleasure!
So the initial issue didn't happen the 2nd time. So if misc_sanity was the
only test
failing then I guess my tests are working fine other than that. Sweet!
When I get a break from work I'll review some patches!
Best,
Ryan
On Sat, Aug 26, 2017 at 12:40 AM, yang...@highgo.com wrote:
> A partition table can be create as bellow:
>
> CREATE TABLE h1 PARTITION OF h;
> CREATE TABLE h2 PARTITION OF h;
> CREATE TABLE h3 PARTITION OF h;
This syntax is very problematic for reasons that have been discussed
on the existing
On Mon, Aug 28, 2017 at 10:16 AM, Alvaro Herrera
wrote:
>> I am fine with however you want to handle it, but it seems odd to me
>> that we don't have a way of embedding INT64_FORMAT in a translatable
>> string. Surely that's going to be a problem in some case, sometime,
>> isn't it?
>
> The way w
Robert Haas wrote:
> On Mon, Aug 28, 2017 at 9:05 AM, Alvaro Herrera
> wrote:
> > Robert Haas wrote:
> >> pg_rewind: Fix some problems when copying files >2GB.
> >
> > I just noticed that this broke pg_rewind translation, because of the
> > INT64_FORMAT marker in the translatable string. The mess
On Mon, Aug 28, 2017 at 9:05 AM, Alvaro Herrera
wrote:
> Robert Haas wrote:
>> pg_rewind: Fix some problems when copying files >2GB.
>
> I just noticed that this broke pg_rewind translation, because of the
> INT64_FORMAT marker in the translatable string. The message catalog now
> has this:
>
> m
On Mon, Aug 28, 2017 at 6:34 PM, Tom Lane wrote:
> Amit Kapila writes:
>> On Mon, Aug 28, 2017 at 6:01 PM, Tom Lane wrote:
>>> Um, what's different about that than before?
>
>> Earlier, we perform the rescan of all the nodes before ExecProcNode,
>> so workers will always start (restart) after th
Ryan Murphy writes:
> I did notice that the test seems to create a ROLE called regress_ecpg_user2:
Right.
> Could that implicitly create a database too? I know that I somehow have a
> database named after my username / postgres role "murftown".
Maybe you've got some tool somewhere that automat
> No, you're reading it backwards: the error is expected, but it's not
> appearing in your results. I can duplicate this if I manually create
> database "regress_ecpg_user2" before running ecpg's installcheck,
> so I guess that's what you did. I can find no evidence that any
> part of the PG regr
On Sat, Aug 26, 2017 at 10:10 AM, yang...@highgo.com
wrote:
> Hi all,
>
> Now we have had the range / list partition, but hash partitioning is not
> implemented yet.
> Attached is a POC patch based on the v10Beta2 to add the
> hash partitioning feature.
> Although we will need more discussions ab
Hello
Looking at your hash partitioning syntax, I implemented a hash partition in a
more concise way, with no need to determine the number of sub-tables, and
dynamically add partitions.
Description
The hash partition's implement is on the basis of the original range / list
partition,and using
Robert Haas wrote:
> pg_rewind: Fix some problems when copying files >2GB.
I just noticed that this broke pg_rewind translation, because of the
INT64_FORMAT marker in the translatable string. The message catalog now
has this:
msgid "received chunk for file \"%s\", offset "
for this source line:
Amit Kapila writes:
> On Mon, Aug 28, 2017 at 6:01 PM, Tom Lane wrote:
>> Um, what's different about that than before?
> Earlier, we perform the rescan of all the nodes before ExecProcNode,
> so workers will always start (restart) after the scan descriptor is
> initialized.
If what you're compl
On Mon, Aug 28, 2017 at 6:01 PM, Tom Lane wrote:
> Amit Kapila writes:
>> With this change, it is quite possible that during rescans workers
>> will not do any work.
>
> Um, what's different about that than before?
>
Earlier, we perform the rescan of all the nodes before ExecProcNode,
so workers
Amit Kapila writes:
> With this change, it is quite possible that during rescans workers
> will not do any work.
Um, what's different about that than before?
regards, tom lane
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your
On Mon, Aug 28, 2017 at 1:59 AM, Tom Lane wrote:
> I wrote:
>> I think that the correct fix probably involves marking each parallel scan
>> plan node as dependent on a pseudo executor parameter, which the parent
>> Gather or GatherMerge node would flag as being changed on each rescan.
>> This woul
Mithun Cy writes:
> I was trying to study NoMovementScanDirection part of heapgettup() and
> heapgettup_pagemode(). If I am right there is no test in test suit to
> hit this code. I did run make check-world could not hit it. Also,
> coverage report in
> https://coverage.postgresql.org/src/backend/
On Tue, Jun 13, 2017 at 4:21 AM, Andres Freund wrote:
> On 2017-06-12 15:12:23 -0400, Robert Haas wrote:
>> Commit 4b4b680c3d6d8485155d4d4bf0a92d3a874b7a65 (Make backend local
>> tracking of buffer pins memory efficient., vintage 2014) seems like a
>> likely culprit here, but I haven't tested.
>
>
On 28 August 2017 at 19:45, Tom Lane wrote:
> Craig Ringer writes:
> > It's a pain having to find the postmaster command line to get the port
> > pg_regress started a server on. We print the port in the pg_regress
> output,
> > why not the socket directory / host?
>
> I'm not following the point
Craig Ringer writes:
> It's a pain having to find the postmaster command line to get the port
> pg_regress started a server on. We print the port in the pg_regress output,
> why not the socket directory / host?
I'm not following the point here. The test postmaster isn't really
going to be around
Ryan Murphy writes:
> And when I look at that diffs file, this is what I see:
> - [NO_PID]: ECPGconnect: could not open database: FATAL: database
> "regress_ecpg_user2" does not exist
> Why would this database not be getting created?
No, you're reading it backwards: the error is expected, but
I don't doubt about a sense of this configuration - but this specific
combination depends on usage - so I don't think so using special option is
good idea.
Although I agree with you that detailed settings are definitely debatable,
I'd say that at least it would be a more reasonable starting p
On Mon, Aug 28, 2017 at 8:02 PM, Kyotaro HORIGUCHI
wrote:
> The first patch (0001-) fixes this problem, preventing the
> problematic state of WAL segments by retarding restart LSN of a
> physical replication slot in a certain condition.
FWIW, I have this patch marked on my list of things to look
Hello,
This problem still occurs on the master.
I rebased this to the current master.
At Mon, 3 Apr 2017 08:38:47 +0900, Michael Paquier
wrote in
> On Mon, Apr 3, 2017 at 7:19 AM, Venkata B Nagothi wrote:
> > As we are already past the commitfest, I am not sure, what should i change
> > the p
On 2017/08/26 3:28, Robert Haas wrote:
> On Mon, Aug 21, 2017 at 2:10 AM, Amit Langote
> wrote:
>> [ new patches ]
>
> I am failing to understand the point of separating PartitionDispatch
> into PartitionDispatch and PartitionTableInfo. That seems like an
> unnecessary multiplication of entities
Craig Ringer wrote:
> On 28 August 2017 at 15:19, Michael Paquier
> wrote:
>
> > On Mon, Aug 28, 2017 at 4:07 PM, Craig Ringer
> > wrote:
> > > == starting postmaster==
> > > running with PID 30235; connect with:
> > > psql "host='/tmp/pg_regress-j74
Hello,
I'll add this to CF2017-09.
At Mon, 06 Mar 2017 18:20:06 +0900 (Tokyo Standard Time), Kyotaro HORIGUCHI
wrote in
<20170306.182006.172683338.horiguchi.kyot...@lab.ntt.co.jp>
> Thank you for the comment.
>
> At Fri, 3 Mar 2017 14:47:20 -0500, Peter Eisentraut
> wrote in
>
> > On 3/1/
2017-08-28 11:05 GMT+02:00 Craig Ringer :
> On 28 August 2017 at 16:23, Fabien COELHO wrote:
>
>>
>> This doesn't really address the original issue though, that it's far from
>>> obvious how to easily and correctly script psql.
>>>
>>
>> That is another interesting argument. I understood that the
Hello,
I'll add this to CF2017-09.
At Tue, 27 Jun 2017 16:27:18 +0900, Amit Langote
wrote in
<75fe42df-b1d8-89ff-596d-d9da0749e...@lab.ntt.co.jp>
> On 2017/06/26 18:44, Kyotaro HORIGUCHI wrote:
> > At Mon, 26 Jun 2017 18:16:42 +0900, Amit Langote wrote:
> >>
> >> I recall I had proposed a fix
Hello,
Currently the message shows the '%d skipped-frozen' message but
it is insufficient to verify the true effect. This is a patch to
show mode as 'aggressive' or 'normal' in the closing message of
vacuum. %d frozen-skipped when 'aggressive mode' shows the true
effect of ALL_FROZEN.
I will add
Thank you for your attention.
At Mon, 14 Aug 2017 17:33:48 -0400, Peter Eisentraut
wrote in
<09fa011f-4536-b05d-0625-11f3625d8...@2ndquadrant.com>
> On 1/24/17 02:58, Kyotaro HORIGUCHI wrote:
> >> BTW, if you set a slightly larger
> >> context size on the patch you might be able to avoid rebase
Craig Ringer wrote:
> On 25 August 2017 at 15:12, Antonin Houska wrote:
>
> How will this play with syslog? csvlog? etc?
There's nothing special about csvlog: the LogStream structure has a
"destination" field, so if particular extension wants this kind of output, it
simply sets the LOG_DESTINA
On 28 August 2017 at 16:23, Fabien COELHO wrote:
>
> This doesn't really address the original issue though, that it's far from
>> obvious how to easily and correctly script psql.
>>
>
> That is another interesting argument. I understood that the issue was
> having to type these options, but now i
Hello Postgres Hackers,
I want to become more helpful to the project, reviewing more patches and
starting to write more of my own - and one of the first steps is to get all
the tests passing so I can confidently review patches. It almost worked...
I typed "make check" and all the tests passed.
T
Hi young,
On Mon, 28 Aug 2017 15:33:46 +0800
"yang...@highgo.com" wrote:
> Hello
>
> Looking at your hash partitioning syntax, I implemented a hash partition in a
> more concise way, with no need to determine the number of sub-tables, and
> dynamically add partitions.
I think it is great wor
This doesn't really address the original issue though, that it's far from
obvious how to easily and correctly script psql.
That is another interesting argument. I understood that the issue was
having to type these options, but now it is also to remember which one are
relevant and wanted, whi
Hello Masahiko-san,
Patch applies cleanly, compiles, works for me.
At least: "Required to invoke" -> "Require to invoke".
Fixed.
I meant the added one about -I, not the pre-existing one about -i.
About the code:
is_no_vacuum should be a bool?
We can change it but I think there is no d
On 28 August 2017 at 15:34, Pavel Stehule wrote:
>
>
> 2017-08-28 9:33 GMT+02:00 Fabien COELHO :
>
>>
>> ISTM that the real pain is the "-v ON_ERRORS_STOP=1" which I occasionally
encountered as well, the other one letter ones are not too bad. Maybe it
would be enough to have a shortcut
On 28 August 2017 at 15:19, Michael Paquier
wrote:
> On Mon, Aug 28, 2017 at 4:07 PM, Craig Ringer
> wrote:
> > == starting postmaster==
> > running with PID 30235; connect with:
> > psql "host='/tmp/pg_regress-j74yFE' port=50848 dbname='regression'"
2017-08-28 9:33 GMT+02:00 Fabien COELHO :
>
> ISTM that the real pain is the "-v ON_ERRORS_STOP=1" which I occasionally
>>> encountered as well, the other one letter ones are not too bad. Maybe it
>>> would be enough to have a shortcut for this one, say "-B"?
>>>
>>
>> I agree with last sentence.
ISTM that the real pain is the "-v ON_ERRORS_STOP=1" which I occasionally
encountered as well, the other one letter ones are not too bad. Maybe it
would be enough to have a shortcut for this one, say "-B"?
I agree with last sentence. I don't think so -qAtX are expected always, but
"-v ON_ERROR
2017-08-28 8:56 GMT+02:00 Fabien COELHO :
>
> I find myself regurgitating the incantation
>>
>> psql -qAtX -v ON_ERRORS_STOP=1
>>
>> quite a bit. It's not ... super friendly.
>>
>> It strikes me that we could possibly benefit from a 'psql --batch' option.
>>
>> Thoughts?
>>
>
> The link between -q
On Mon, Aug 28, 2017 at 4:07 PM, Craig Ringer wrote:
> == starting postmaster==
> running with PID 30235; connect with:
> psql "host='/tmp/pg_regress-j74yFE' port=50848 dbname='regression'"
> == creating database "regression" ==
== starting postmaster==
running with PID 30235; connect with:
psql "host='/tmp/pg_regress-j74yFE' port=50848 dbname='regression'"
== creating database "regression" ==
On 28 August 2017 at 14:08, Michael Paquier
wrote:
On 28 August 2017 at 14:56, Fabien COELHO wrote:
>
> I find myself regurgitating the incantation
>>
>> psql -qAtX -v ON_ERRORS_STOP=1
>>
>> quite a bit. It's not ... super friendly.
>>
>> It strikes me that we could possibly benefit from a 'psql --batch' option.
>>
>> Thoughts?
>>
>
> The link be
80 matches
Mail list logo