On Mon, Aug 07, 2023 at 09:24:02AM +0530, Amit Kapila wrote:
>
> I think autovacuum is not enabled during the upgrade. See comment "Use
> -b to disable autovacuum." in start_postmaster(). However, I am not
> sure if there can't be any additional WAL from checkpointer or
> bgwriter. Checkpointer
At Fri, 4 Aug 2023 17:07:51 +0200, "Drouvot, Bertrand"
wrote in
> SynRep currently appears in "IPC" and "LWLock" (see [2])
> WALWrite currently appears in "IO" and "LWLock" (see [2])
>
> I think that can lead to confusion and it would be better to avoid
> duplicate wait event
> name across
hi.
based on Heikki v3.
I made some changes:
array_in: dim[6] all initialize with -1, lBound[6] all initialize with 1.
if ReadArrayDimensions called, then corresponding dimension lBound
will replace the initialized default 1 value.
ReadArrayStr, since array_in main function initialized dim array,
On Wed, Aug 2, 2023 at 7:46 AM Jonathan S. Katz wrote:
>
> Can I take this a step further on the user interface and ask why the
> flag would be "--include-logical-replication-slots" vs. being enabled by
> default?
>
> Are there reasons why we wouldn't enable this feature by default on
>
On Sun, Aug 6, 2023 at 7:54 PM José Neves wrote:
>
> A follow-up on this. Indeed, a new commit-based approach solved my missing
> data issues.
> But, getting back to the previous examples, how are server times expected to
> be logged for the xlogs containing these records?
>
I think it should
After commit 7389aad6, I think commit 8acd8f86's linker changes (+
meson.build's equivalent) must now be redundant?
On Sun, Aug 6, 2023 at 6:02 PM Masahiko Sawada wrote:
>
> On Wed, Aug 2, 2023 at 5:13 PM Hayato Kuroda (Fujitsu)
> wrote:
> >
> > > 4.
> > > + /*
> > > + * Check that all logical replication slots have reached the current WAL
> > > + * position.
> > > + */
> > > + res = executeQueryOrDie(conn,
>
Hi:
> For all the people who are interested in this topic, I will post a
> planner support function soon, you can check that then.
>
>
The updated patch doesn't need users to change their codes and can get
better performance. Thanks for all the feedback which makes things better.
To verify
On Mon, 7 Aug 2023 11:02:48 +0900
Yugo NAGATA wrote:
> On Sat, 05 Aug 2023 12:16:11 +0900 (JST)
> Tatsuo Ishii wrote:
>
> > > Hi,
> > >
> > > I would like to propose to add an option to pgbench so that benchmark
> > > can quit immediately when any client is aborted. Currently, when a
> > >
On Sat, 05 Aug 2023 12:16:11 +0900 (JST)
Tatsuo Ishii wrote:
> > Hi,
> >
> > I would like to propose to add an option to pgbench so that benchmark
> > can quit immediately when any client is aborted. Currently, when a
> > client is aborted due to some error, for example, network trouble,
> >
Thomas Munro writes:
> On Mon, Aug 7, 2023 at 7:21 AM Konstantin Knizhnik wrote:
>> Two tests are failed because of sync scan - this tests cluster.sql and
>> portals.sql perform seqscan without explicit order by and expect that
>> data will be returned in particular order. But because of sync
On Sun, Aug 6, 2023 at 3:28 PM Peter Geoghegan wrote:
> I decided to verify my understanding by checking what would happen
> when I ran the OR-heavy tenk1 regression test query against a
> combination of your patch, and v7 of the OR-to-SAOP transformation
> patch. (To be clear, this is without my
On Mon, Aug 7, 2023 at 7:21 AM Konstantin Knizhnik wrote:
> Two tests are failed because of sync scan - this tests cluster.sql and
> portals.sql perform seqscan without explicit order by and expect that
> data will be returned in particular order. But because of sync scan it
> doesn't happen.
Konstantin Knizhnik writes:
> Is it is ok, that regression tests do not pass with small value of
> shared buffers (for example 1Mb)?
There are quite a few GUC settings with which you can break the
regression tests. I'm not especially bothered by this one.
> More general question - is it
On Sun, Aug 6, 2023 at 1:13 PM Peter Geoghegan wrote:
> Since you're not relying on the nbtree work at all here, really (just
> on the transformation process itself), the strategic risk that this
> adds to your project isn't too great. It's not like this ties the
> success of your patch to the
On Mon, Aug 7, 2023 at 8:40 AM Christoph Berg wrote:
> 2023-08-06 17:21:24.078 UTC [127] 031_recovery_conflict.pl FATAL:
> unrecognized conflict mode: 7
Thanks for testing! Would you mind trying v8 from that thread? V7
had a silly bug (I accidentally deleted a 'case' label while cleaning
Hi,
## Laetitia Avrot (laetitia.av...@gmail.com):
> I understand your point and sure enough, my customer could set and use the
> cluster_name for that purpose. I totally disagree with using
> inet_server_addr() for that purpose as there are so many different network
> settings with VIPs and so
On Sat, Aug 5, 2023 at 7:01 PM Peter Geoghegan wrote:
> Of course this immediately makes me wonder: shouldn't your patch be
> able to perform an additional transformation here? You know, by
> transforming "a.x = 42 OR a.x = 44" into "a IN (42, 44)"? Although I
> haven't checked for myself, I
Re: Thomas Munro
> It's great that you can reproduce this semi-reliably! I've rebased
> the patch, hoping you can try it out.
Unfortunately very semi, today I didn't get to the same point where it
exited after test 7, but got some other timeouts. Not even sure they
are related to this (?)
On Sun, Aug 6, 2023 at 10:23 AM Tomas Vondra
wrote:
> > The index AM is entitled to make certain assumptions of opclass
> > members -- assumptions that cannot be made during expression
> > evaluation.
> Thanks for reminding me, I keep forgetting about this.
I was almost certain that you already
Hi,
On 2023-08-02 16:51:29 -0700, Matt Smiley wrote:
> I thought it might be helpful to share some more details from one of the
> case studies behind Nik's suggestion.
>
> Bursty contention on lock_manager lwlocks recently became a recurring cause
> of query throughput drops for GitLab.com, and
Hi all,
I have addressed the pg version compatibility with the FORCE option in
drop. Here is the last version of the patch
On Tue, Aug 1, 2023 at 6:19 PM Ahmed Ibrahim
wrote:
> Hi Gurjeet,
>
> I have addressed all your comments except for the tests.
>
> I have tried adding test cases but I
Hi hackers,
Is it is ok, that regression tests do not pass with small value of
shared buffers (for example 1Mb)?
Two tests are failed because of sync scan - this tests cluster.sql and
portals.sql perform seqscan without explicit order by and expect that
data will be returned in particular
On 8/5/23 02:53, Peter Geoghegan wrote:
> ...
>
>> Right. This however begs a question - why would we actually need to
>> check the visibility map before evaluating the index filter, when the
>> index tuple alone is clearly good enough for the bitmapOr plan?
>>
>> Because if we didn't need to do
On 8/3/23 22:39, Tomas Vondra wrote:
> On 8/3/23 01:51, Matt Smiley wrote:
>> I thought it might be helpful to share some more details from one of the
>> case studies behind Nik's suggestion.
>>
>> Bursty contention on lock_manager lwlocks recently became a recurring
>> cause of query
Hi there, hope to find you all well.
A follow-up on this. Indeed, a new commit-based approach solved my missing data
issues.
But, getting back to the previous examples, how are server times expected to be
logged for the xlogs containing these records?
With these 2 transactions:
T-1
INSERT
On Wed, Aug 2, 2023 at 5:13 PM Hayato Kuroda (Fujitsu)
wrote:
>
> > 4.
> > + /*
> > + * Check that all logical replication slots have reached the current WAL
> > + * position.
> > + */
> > + res = executeQueryOrDie(conn,
> > + "SELECT slot_name FROM pg_catalog.pg_replication_slots "
> > + "WHERE
On Fri, 4 Aug 2023 at 16:05, Richard Guo wrote:
>
>
> On Fri, Aug 4, 2023 at 10:36 AM David Rowley wrote:
>>
>> The whole lnext() stuff all feels a bit old now that Lists are arrays.
>> I think we'd be better adjusting the code to pass the List index where
>> we start from rather than the
On Sun, Aug 06, 2023 at 01:37:42PM +0800, jacktby jacktby wrote:
>
> I need to build a new datatype. It can contains different datatypes, like
> ‘(1,’a’,2.0)’,it’s a (interger,string,float) tuple type
Is there any reason why you can't simply rely on the record datatype?
> and Then I need to
>
29 matches
Mail list logo
