Hello Andres,
- Move fsync as early as possible, suggested by Andres Freund?
My opinion is that this should be left out for the nonce.
"for the nonce" - what does that mean?
Nonce \Nonce\ (n[o^]ns), n. [For the nonce, OE. for the nones, ...
{for the nonce}, i. e. for the present time
On Wed, Jun 17, 2015 at 10:52 AM, Tomas Vondra wrote:
> Hi,
>
> I'm currently running some tests on a 3TB TPC-H data set, and I tripped
> over a pretty bad n_distinct underestimate, causing OOM in HashAgg (which
> somehow illustrates the importance of the memory-bounded hashagg patch Jeff
> Davis
Alvaro Herrera wrote:
> Tom Lane wrote:
> > Andres Freund writes:
>
> > > Uh. I'm pretty sure there were some back when that patch went in. And
> > > there definitely used to be a couple earlier. I guess itanium really is
> > > dying (mixed bad: It's a horrible architecture, but more coverage wou
Has anybody noticed the way castoroides is randomly failing?
SELECT test_shm_mq_pipelined(16384, (select
string_agg(chr(32+(random()*95)::int), '') from generate_series(1,27)),
200, 3);
! PANIC: stuck spinlock (100cb92f4) detected at atomics.c:30
! server closed the connection unexpectedl
On Fri, Jun 19, 2015 at 1:39 PM, Tomas Vondra
wrote:
> On 06/19/2015 09:48 PM, Jeff Janes wrote:
>
>> On Fri, Jun 19, 2015 at 12:27 PM, Tomas Vondra
>> mailto:tomas.von...@2ndquadrant.com>>
>> wrote:
>>
>> But I think you might be on to something, because I manually
>> collected a random
I came across the following bug this week:
Session 0:
begin;
create schema bug;
alter default privileges in schema bug grant all on tables to postgres;
commit;
Session 1:
begin;
alter default privileges in schema bug grant all on tables to postgres;
Session 2:
alter default privileges in schema
On Fri, Jun 19, 2015 at 11:45 PM, Tom Lane wrote:
> Michael Paquier writes:
>> Attached is a patch fixing those problems and improving the log
>> facility as it really helped me out with those issues. The simplest
>> fix would be to include the -w switch missing in the tests of
>> pg_rewind and p
Jim Nasby wrote:
> On 6/14/15 9:50 AM, Alvaro Herrera wrote:
> >+values[0] = MultiXactState->oldestMultiXactId;
>
> What about oldestOffset and offsetStopLimit? Seems those would be useful
> too. Looks good other than that.
Yeah, that's what I was trying to say. How about this?
I realized t
On Sat, Jun 20, 2015 at 12:44 AM, Tom Lane wrote:
> Andres Freund writes:
>> On 2015-06-19 11:16:18 -0400, Robert Haas wrote:
>>> On Fri, Jun 19, 2015 at 11:07 AM, Tom Lane wrote:
I wonder whether it's such a good idea for the postmaster to give
up waiting before all children are gone
On Sat, Jun 20, 2015 at 12:07 AM, Tom Lane wrote:
> Michael Paquier writes:
>> Now if we look at RewindTest.pm, there is the following code:
>> if ($test_master_datadir)
>> {
>> system
>> "pg_ctl -D $test_master_datadir -s -m immediate stop
>> 2>
On 06/19/2015 09:48 PM, Jeff Janes wrote:
On Fri, Jun 19, 2015 at 12:27 PM, Tomas Vondra
mailto:tomas.von...@2ndquadrant.com>> wrote:
But I think you might be on to something, because I manually
collected a random sample with 30k rows (by explicitly generating
30k random TIDs), and I
On Fri, 19 Jun 2015 at 21:05 Petr Jelinek wrote:
> On 2015-06-19 09:08, Brendan Jurd wrote:
> > I
> > think it would be convenient and user-friendly to complete the opening
> > bracket -- it would make it perfectly clear that an argument is required
> > for the syntax to be valid.
> >
>
> Agreed,
Hi,
On 2015-06-11 00:15:21 -0400, Bruce Momjian wrote:
> I have committed the first draft of the 9.5 release notes. You can view
> the output here:
I'm looking through all the commits, checking which I think should
possibly be mentioned additionally:
- 9f03ca91 - Speed up CREATE INDEX by avoidin
On Fri, Jun 19, 2015 at 12:27 PM, Tomas Vondra wrote:
> But I think you might be on to something, because I manually collected a
> random sample with 30k rows (by explicitly generating 30k random TIDs), and
> I get this:
>
> tpch=# select cnt, count(*) from (select l_orderkey, count(*) AS cnt fr
On 06/19/2015 08:32 PM, Jeff Janes wrote:
On Wed, Jun 17, 2015 at 10:52 AM, Tomas Vondra
mailto:tomas.von...@2ndquadrant.com>> wrote:
Hi,
I'm currently running some tests on a 3TB TPC-H data set, and I
tripped over a pretty bad n_distinct underestimate, causing OOM in
HashAgg
On 06/19/2015 02:51 PM, Tom Lane wrote:
> Josh Berkus writes:
>> On 05/14/2015 12:10 PM, Fabien COELHO wrote:
>>> Add backslash continuations to pgbench custom scripts.
>
>> I don't personally agree. I believe that it it worth breaking backwards
>> compatibility to support line breaks in pgbench
Tom Lane wrote:
> Andres Freund writes:
> > On 2015-06-19 13:56:21 -0300, Alvaro Herrera wrote:
> >> We discussed this when that patch got in (82233ce7ea42d6b). The reason
> >> for not waiting, it was argued, is that the most likely reason for those
> >> processes not to have already gone away by
On 06/19/2015 02:33 AM, Michael Paquier wrote:
Hi all,
In pg_regress.c, there is the following code:
static void
stop_postmaster(void)
{
if (postmaster_running)
{
/* We use pg_ctl to issue the kill and wait for stop */
charbuf[MAXP
Josh Berkus writes:
> On 05/14/2015 12:10 PM, Fabien COELHO wrote:
>> Add backslash continuations to pgbench custom scripts.
> I don't personally agree. I believe that it it worth breaking backwards
> compatibility to support line breaks in pgbench statements, and that if
> we're not going to do
On Wed, Jun 17, 2015 at 10:52 AM, Tomas Vondra wrote:
> Hi,
>
> I'm currently running some tests on a 3TB TPC-H data set, and I tripped
> over a pretty bad n_distinct underestimate, causing OOM in HashAgg (which
> somehow illustrates the importance of the memory-bounded hashagg patch Jeff
> Davis
On 05/14/2015 12:10 PM, Fabien COELHO wrote:
>
> Add backslash continuations to pgbench custom scripts.
>
> The benefit of this approach is that it is upward compatible, and it is
> also pretty simple to implement. The downside is that backslash
> continuation is not the best syntax ever invented
Andres Freund writes:
> On 2015-06-19 13:56:21 -0300, Alvaro Herrera wrote:
>> We discussed this when that patch got in (82233ce7ea42d6b). The reason
>> for not waiting, it was argued, is that the most likely reason for those
>> processes not to have already gone away by the time we send SIGKILL
Hi,
On 2015-06-11 00:15:21 -0400, Bruce Momjian wrote:
> I have committed the first draft of the 9.5 release notes. You can view
> the output here:
So, I did a pass through master's state:
>
>
>Add Block Range Indexes (BRIN)
>(Álvaro Herrera, Heikki Linnakangas, Emr
Alvaro Herrera wrote:
> We discussed this when that patch got in (82233ce7ea42d6b). The reason
> for not waiting, it was argued, is that the most likely reason for those
> processes not to have already gone away by the time we send SIGKILL was
> that they are stuck somewhere in the kernel, and so
On 2015-06-19 13:56:21 -0300, Alvaro Herrera wrote:
> We discussed this when that patch got in (82233ce7ea42d6b). The reason
> for not waiting, it was argued, is that the most likely reason for those
> processes not to have already gone away by the time we send SIGKILL was
> that they are stuck so
On 2015-06-17 18:10:42 -0300, Alvaro Herrera wrote:
> Yeah, the case is pretty weird and I'm not really sure that the server
> ought to be expected to behave. But if this is actually the only part
> of the server that misbehaves because of sudden gigantic time jumps, I
> think it's fair to patch i
Tom Lane wrote:
> Andres Freund writes:
> > On 2015-06-19 11:16:18 -0400, Robert Haas wrote:
> >> On Fri, Jun 19, 2015 at 11:07 AM, Tom Lane wrote:
> >>> I wonder whether it's such a good idea for the postmaster to give
> >>> up waiting before all children are gone (postmaster.c:1722 in HEAD).
>
Hi list
One of my databases failed to upgrade successfully and produced this error
in the copying phase:
error while copying relation "pg_catalog.pg_largeobject"
("/srv/ssd/PG_9.3_201306121/1/12023" to "/PG_9.4_201409291/1/12130"): No
such file or directory
Turns out this happens when either the
Le 19/06/2015 17:21, Andres Freund a écrit :
> On 2015-06-19 11:19:23 -0400, Magnus Hagander wrote:
>> On Fri, Jun 19, 2015 at 11:17 AM, Stephen Frost wrote:
>>
>>> * Andres Freund (and...@anarazel.de) wrote:
To make slot usage in pg_receivexlog easier, should we add
--create-slot-if-n
Tom Lane wrote:
> Alvaro Herrera writes:
> > Yeah, the case is pretty weird and I'm not really sure that the server
> > ought to be expected to behave. But if this is actually the only part
> > of the server that misbehaves because of sudden gigantic time jumps, I
> > think it's fair to patch it.
I think we're missing a dependency declaration from the backend to
libpgport in the makefiles somewhere. I just git updated on 9.1 on a
tree that was already built, ran make, and got this:
../../src/port/libpgport_srv.a(dirmod_srv.o): In function `rmtree':
/pgsql/source/REL9_1_STABLE/src/port/dir
During the Developer Unconference at PGCon, we made this wiki page to
track the MultiXact bugs that are outstanding:
https://wiki.postgresql.org/wiki/MultiXact_Bugs
Let's update this page as we find or fix things, and let's release
when we've fixed enough stuff.
Even if you are not involved in t
Andres Freund writes:
> On 2015-06-19 11:16:18 -0400, Robert Haas wrote:
>> On Fri, Jun 19, 2015 at 11:07 AM, Tom Lane wrote:
>>> I wonder whether it's such a good idea for the postmaster to give
>>> up waiting before all children are gone (postmaster.c:1722 in HEAD).
>> I doubt it.
> Seconded.
Petr Jelinek writes:
> On 2015-06-19 01:04, Petr Jelinek wrote:
>> On 2015-06-19 00:38, Petr Jelinek wrote:
>>> On 2015-06-18 22:04, Tom Lane wrote:
By the by, the tablesample additions to range_table_mutator are
obviously broken.
> Apparently it's not a good idea to do this at 1AM afte
Hi,
On 2015-06-17 08:24:38 +0200, Fabien COELHO wrote:
> Here is version 3, including many performance tests with various settings,
> representing about 100 hours of pgbench run. This patch aims at improving
> checkpoint I/O behavior so that tps throughput is improved, late
> transactions are less
On 2015-06-19 11:19:23 -0400, Magnus Hagander wrote:
> On Fri, Jun 19, 2015 at 11:17 AM, Stephen Frost wrote:
>
> > * Andres Freund (and...@anarazel.de) wrote:
> > > To make slot usage in pg_receivexlog easier, should we add
> > > --create-slot-if-not-exists? That'd mean you could run the same com
> To make slot usage in pg_receivexlog easier, should we add
> --create-slot-if-not-exists? That'd mean you could run the same command
> the first and later invocation.
+1 (with a shorter name please, if you can find one... )
--
Cédric Villemain +33 (0)6 20 30 22 52
http://2ndQuadrant.fr/
Postgr
On 2015-06-19 11:16:18 -0400, Robert Haas wrote:
> On Fri, Jun 19, 2015 at 11:07 AM, Tom Lane wrote:
> > I wonder whether it's such a good idea for the postmaster to give
> > up waiting before all children are gone (postmaster.c:1722 in HEAD).
>
> I doubt it.
Seconded. It's pretty bad to possibl
On Fri, Jun 19, 2015 at 11:17 AM, Stephen Frost wrote:
> * Andres Freund (and...@anarazel.de) wrote:
> > To make slot usage in pg_receivexlog easier, should we add
> > --create-slot-if-not-exists? That'd mean you could run the same command
> > the first and later invocation.
>
> Yes, please.
>
* Andres Freund (and...@anarazel.de) wrote:
> To make slot usage in pg_receivexlog easier, should we add
> --create-slot-if-not-exists? That'd mean you could run the same command
> the first and later invocation.
Yes, please.
Thanks!
Stephen
signature.asc
Description: D
On Fri, Jun 19, 2015 at 11:07 AM, Tom Lane wrote:
> I wonder whether it's such a good idea for the postmaster to give
> up waiting before all children are gone (postmaster.c:1722 in HEAD).
I doubt it.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
-
On Fri, Jun 19, 2015 at 5:52 AM, Jeevan Chalke
wrote:
> I have observed that we are not tab-completing word PASSWORD in the
> following
> syntaxes:
>
> 1.
> CREATE|ALTER ROLE|USER rolname
>
> 2.
> CREATE|ALTER ROLE|USER rolname WITH
>
> PASSWORD is used many times and should be in the tab-complete
Michael Paquier writes:
> Now if we look at RewindTest.pm, there is the following code:
> if ($test_master_datadir)
> {
> system
> "pg_ctl -D $test_master_datadir -s -m immediate stop
> 2> /dev/null";
> }
> if ($test_standby_datadir
Michael Paquier writes:
> Except if I am missing something, the comment mentions that we should
> wait for the stop but there is no -w switch in the command of pg_ctl.
> Doesn't the patch attached make sense?
Per pg_ctl's help output:
(The default is to wait for shutdown, but not for start or re
On Fri, Jun 19, 2015 at 2:33 AM, Michael Paquier
wrote:
> In pg_regress.c, there is the following code:
> static void
> stop_postmaster(void)
> {
> if (postmaster_running)
> {
> /* We use pg_ctl to issue the kill and wait for stop */
> char
Hi,
To make slot usage in pg_receivexlog easier, should we add
--create-slot-if-not-exists? That'd mean you could run the same command
the first and later invocation.
Greetings,
Andres Freund
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscri
Michael Paquier writes:
> Attached is a patch fixing those problems and improving the log
> facility as it really helped me out with those issues. The simplest
> fix would be to include the -w switch missing in the tests of
> pg_rewind and pg_ctl though.
I agree with adding the -w switch and will
Robert Haas writes:
> On Sun, Jun 14, 2015 at 11:34 PM, Craig Ringer wrote:
>> I frequently see users confused by one of our more common and less
>> clear error messages:
>>
>> fe_sendauth: no password supplied
> I've never seen this error message, but I'm not opposed to improving
> it in some
On Sun, Jun 14, 2015 at 11:34 PM, Craig Ringer wrote:
> I frequently see users confused by one of our more common and less
> clear error messages:
>
> fe_sendauth: no password supplied
I've never seen this error message, but I'm not opposed to improving
it in some way.
--
Robert Haas
Enterp
On Wed, Jun 17, 2015 at 4:11 PM, Jim Nasby wrote:
> On 6/16/15 8:26 AM, Sawada Masahiko wrote:
>>
>> I noticed while using gin function of pageinspect that there are some
>> inconsistency data types.
>> For example, data type of GinMetaPageData.head, and tail is
>> BlockNumber, i.g, uint32.
>> Bu
On Thu, Jun 11, 2015 at 9:32 PM, Qingqing Zhou
wrote:
> On Fri, Jun 5, 2015 at 10:59 AM, Tom Lane wrote:
>> So I really doubt that anyone would have any enthusiasm for saddling btree
>> with a similar mechanism. It's complicated (and has been the cause of
>> multiple bugs); it's hard to figure o
On 19 June 2015 at 14:30, Robert Haas wrote:
> > So I really doubt that anyone would have any enthusiasm for saddling
> btree
> > with a similar mechanism. It's complicated (and has been the cause of
> > multiple bugs); it's hard to figure out when is the optimal time to flush
> > the pending in
> So I really doubt that anyone would have any enthusiasm for saddling btree
> with a similar mechanism. It's complicated (and has been the cause of
> multiple bugs); it's hard to figure out when is the optimal time to flush
> the pending insertions; and it slows down searches in favor of making
>
On Fri, Jun 19, 2015 at 8:18 AM, Robert Haas wrote:
> On Fri, Jun 19, 2015 at 12:14 AM, Michael Paquier
> wrote:
>>> Listing the directories with pg_ls_dir() has the same problem.
>>
>> (After some discussion on IM with Heikki on this one).
>> This is actually more tricky because pg_ls_dir() does
On Fri, Jun 19, 2015 at 12:14 AM, Michael Paquier
wrote:
>> Listing the directories with pg_ls_dir() has the same problem.
>
> (After some discussion on IM with Heikki on this one).
> This is actually more tricky because pg_ls_dir() does not return '.'
> or '..' that we could use to identify that
On 2015-06-19 09:08, Brendan Jurd wrote:
On Sun, 14 Jun 2015 at 20:44 Petr Jelinek mailto:p...@2ndquadrant.com>> wrote:
looks like I omitted psql tab completion from the TABLESAMPLE patch. The
attached patch adds it.
Hi Petr,
I'm doing an initial review of this patch.
Thanks
With
On Thu, Jun 18, 2015 at 3:52 PM, Michael Paquier wrote:
> I think that it would be useful as well to improve the buildfarm
> output. Thoughts?
And after running the tests more or less 6~7 times in a row on a PI, I
have been able to trigger the problem and I think that I have found
its origin. Firs
Hi,
I have observed that we are not tab-completing word PASSWORD in the
following
syntaxes:
1.
CREATE|ALTER ROLE|USER rolname
2.
CREATE|ALTER ROLE|USER rolname WITH
PASSWORD is used many times and should be in the tab-complete list.
Was there any reason we have deliberately kept this out?
If y
Hi, andres
Thank you for your comment, and sorry for late response.
> The question is how to actually get useful estimates. As there's no
> progress report for indvidiual COPY and CREATE INDEX commands you'll, in
> many cases, have very irregular progress updates. In many many cases
> most of the
On Sun, 14 Jun 2015 at 20:44 Petr Jelinek wrote:
> looks like I omitted psql tab completion from the TABLESAMPLE patch. The
> attached patch adds it.
>
Hi Petr,
I'm doing an initial review of this patch.
It applies and compiles cleanly. Code style is consistent with its
surroundings.
With th
60 matches
Mail list logo