On Thu, Jun 4, 2020 at 5:37 AM Thomas Munro wrote:
> On Thu, Jun 4, 2020 at 1:53 PM Thomas Munro
> wrote:
> > On Thu, Jun 4, 2020 at 1:35 PM Tom Lane wrote:
> > > Ah, it's better if I put the pqReadData call into *both* the paths
> > > where 1f39a1c06 made pqSendSome give up. The attached patc
On Wed, 2020-06-03 at 19:57 -0400, Chapman Flack wrote:
> Ok, so a person in the situation described here, who is not in a position
> to demand changes in an organizational policy (whether or not it seems
> ill-conceived to you or even to him/her), is facing this question:
>
> What are the "safest
Hello, Andrey.
At Wed, 3 Jun 2020 15:00:06 +0500, Andrey Lepikhov
wrote in
> This patch no longer applies cleanly.
> In addition, code comments contain spelling errors.
Sure. Thaks for noticing of them and sorry for the many typos.
Additional item in WaitEventIPC conflicted with this.
I foun
On Wed, Jun 03, 2020 at 02:45:50PM -0400, Tom Lane wrote:
> In the abstract, I agree with Peter's point that we shouldn't alter
> user-given strings without need. However, I think there's strong
> reason for canonicalizing the data directory and config file locations.
> We access those both before
On Thu, Jun 4, 2020 at 9:10 AM Andres Freund wrote:
>
> Hi,
>
> On 2020-06-04 08:10:07 +0530, Amit Kapila wrote:
> > On Thu, Jun 4, 2020 at 12:09 AM Andres Freund wrote:
> > > > I strongly disagree with the idea of "just sync(ing) it up at the end
> > > > of parallelism". That seems like a comple
> ...
>> Umm, did you see any indication that they intend to allow "-" /anywhere/
>> in a time interval (with the exception of between year and month, month
>> and day in the alternate form, as simple delimiters, not as minus?
>> (Maybe you did; I'm looking at a publicly-accessible 2016 draft.)
>
>
On Wed, Jun 3, 2020 at 9:46 PM, Tom Lane wrote:
> ...
> ISTM that if the standard intended to allow that, it'd be pretty
> clear. I looked through the 8601 spec just now, and I can't see any
> indication whatever that they intend to allow "-" before P.
To be fair, I do not have an access to 20
Chapman Flack writes:
> On 06/03/20 22:46, Tom Lane wrote:
>> "Isn't quite clear"? ISTM that if the standard intended to allow that,
>> it'd be pretty clear. I looked through the 8601 spec just now, and
>> I can't see any indication whatever that they intend to allow "-" before P.
> Umm, did yo
On Wed, Jun 03, 2020 at 02:40:54PM +0200, Daniel Gustafsson wrote:
> Sounds good, thanks!
Okay, done then.
--
Michael
signature.asc
Description: PGP signature
On Thu, Jun 4, 2020 at 3:36 PM Thomas Munro wrote:
> Here's what I tested.
In passing, I noticed that this:
$ psql ...
psql: error: could not connect to server: private key file
"src/test/ssl/ssl/client-revoked.key" has group or world access;
permissions should be u=rw (0600) or less
... leads
On 06/03/20 22:46, Tom Lane wrote:
> "Isn't quite clear"? ISTM that if the standard intended to allow that,
> it'd be pretty clear. I looked through the 8601 spec just now, and
> I can't see any indication whatever that they intend to allow "-" before P.
Umm, did you see any indication that they
On Wed, Jun 3, 2020 at 12:02 PM Masahiko Sawada
wrote:
>
> On Wed, 3 Jun 2020 at 14:50, Amit Kapila wrote:
> >
> >
> > If the intention is to keep the first version simple, then why do we
> > want to support any mode other than 'required'? I think it will limit
> > its usage for the cases where
Hi,
On 2020-06-04 08:10:07 +0530, Amit Kapila wrote:
> On Thu, Jun 4, 2020 at 12:09 AM Andres Freund wrote:
> > > I strongly disagree with the idea of "just sync(ing) it up at the end
> > > of parallelism". That seems like a completely unprincipled approach to
> > > the problem. Either the comman
On Thu, Jun 4, 2020 at 1:53 PM Thomas Munro wrote:
> On Thu, Jun 4, 2020 at 1:35 PM Tom Lane wrote:
> > Ah, it's better if I put the pqReadData call into *both* the paths
> > where 1f39a1c06 made pqSendSome give up. The attached patch seems
> > to fix the issue for the "pgbench -i" scenario, wit
Mikhail Titov writes:
> I'd like to propose a simple patch to allow for negative ISO 8601
> intervals with leading minus, e.g. -PT1H besides PT-1H. It seems that
> standard isn't quite clear on negative duration.
"Isn't quite clear"? ISTM that if the standard intended to allow that,
it'd be pret
On Thu, Jun 4, 2020 at 12:09 AM Andres Freund wrote:
>
> Hi,
>
> On 2020-06-03 12:13:14 -0400, Robert Haas wrote:
> > On Mon, May 18, 2020 at 12:48 AM Amit Kapila
> > wrote:
> > > In the above case, even though we are executing a single command from
> > > the user perspective, but the currentCom
On Wed, Jun 03, 2020 at 12:40:38PM -0300, Euler Taveira wrote:
> On Wed, 3 Jun 2020 at 03:54, Michael Paquier wrote:
>> I have bumped into $subject, causing a replica identity index to
>> be considered as dropped if running REINDEX CONCURRENTLY on it. This
>> means that the old tuple information
On 2020-Jun-03, Andres Freund wrote:
> On 2020-06-03 18:27:12 -0400, Alvaro Herrera wrote:
> > On 2020-Jun-03, Andres Freund wrote:
> > > I don't think we should prohibit this. For one, it'd probably break some
> > > clients, without a meaningful need.
> >
> > There *is* a need, namely to keep co
On Wed, Jun 03, 2020 at 06:33:11PM -0700, Andres Freund wrote:
> On 2020-06-03 18:27:12 -0400, Alvaro Herrera wrote:
>> There *is* a need, namely to keep complexity down. This is quite
>> convoluted, it's got a lot of historical baggage because of the way it
>> was implemented, and it's very diffi
Michael Paquier writes:
> On Wed, Jun 03, 2020 at 12:36:34AM -0400, Tom Lane wrote:
>> Should we think about adding automated detection of this type of
>> mistake? I don't like the attached as-is because of the #include
>> footprint expansion, but maybe we can find a better way.
> I think that t
On Thu, Jun 4, 2020 at 1:35 PM Tom Lane wrote:
> I wrote:
> > * pqSendSome() is responsible not only for pushing out data, but for
> > calling pqReadData in any situation where it can't get rid of the data
> > promptly. 1f39a1c06 overlooked that requirement, and the upshot is
> > that we don't ne
On Wed, Jun 03, 2020 at 12:36:34AM -0400, Tom Lane wrote:
> Should we think about adding automated detection of this type of
> mistake? I don't like the attached as-is because of the #include
> footprint expansion, but maybe we can find a better way.
I think that this one first boils down to the
I wrote:
> * pqSendSome() is responsible not only for pushing out data, but for
> calling pqReadData in any situation where it can't get rid of the data
> promptly. 1f39a1c06 overlooked that requirement, and the upshot is
> that we don't necessarily notice that the connection is broken (it's
> pqR
Hi,
On 2020-06-03 18:27:12 -0400, Alvaro Herrera wrote:
> On 2020-Jun-03, Andres Freund wrote:
> > I don't think we should prohibit this. For one, it'd probably break some
> > clients, without a meaningful need.
>
> There *is* a need, namely to keep complexity down. This is quite
> convoluted, i
On 2020-Jun-03, Mark Dilger wrote:
> The name "relkind" normally refers to a field of type 'char' with
> values like 'r' for "table" and 'i' for "index". In AlterTableStmt
> and CreateTableAsStmt, this naming convention was abused for a field
> of type enum ObjectType.
I agree that "relkind" her
On 2020-Jun-02, Michael Paquier wrote:
> I can note as well that StartLogicalReplication() moves in this sense
> by setting xlogreader to be the one from logical_decoding_ctx once the
> decoding context has been created.
>
> This results in the attached. The extra test from upthread to check
> t
On 2020-Jun-03, Andres Freund wrote:
> I don't think we should prohibit this. For one, it'd probably break some
> clients, without a meaningful need.
There *is* a need, namely to keep complexity down. This is quite
convoluted, it's got a lot of historical baggage because of the way it
was implem
On 06/03/20 08:07, Ants Aasma wrote:
> I think the "why" the org cert is not root was already made clear, that is
> the copmany policy.
Thank you, yes, that was what I had intended to convey, and you have saved
me finishing a weedsier follow-up message hoping to convey it better.
> I don't think
On Wed, Jun 3, 2020 at 3:18 PM Soumyadeep Chakraborty
wrote:
> Idk if that is a lesser evil than the workers
> being idle..probably not?
Apologies, I meant that the extra atomic fetches is probably a lesser
evil than the workers being idle.
Soumyadeep
Andres Freund writes:
> When libpq is used to COPY data to the server, it doesn't properly
> handle errors.
> This is partially an old problem, and partially got recently
> worse. Before the below commit we detected terminated connections, but
> we didn't handle copy failing.
Yeah. After poking
On Sat, May 23, 2020 at 12:00 AM Robert Haas
wrote:
> I think there's a significant difference. The idea I remember being
> discussed at the time was to divide the relation into equal parts at
> the very start and give one part to each worker. I think that carries
> a lot of risk of some workers f
> It doesn't look like it's using table_block_parallelscan_nextpage() as
> a block allocator so it's not affected by the patch. It has its own
> thing zs_parallelscan_nextrange(), which does
> pg_atomic_fetch_add_u64(&pzscan->pzs_allocatedtids,
> ZS_PARALLEL_CHUNK_SIZE), and that macro is 0x10
Hi,
On 2020-06-02 14:23:50 +0900, Fujii Masao wrote:
> On 2020/06/02 13:24, Michael Paquier wrote:
> > On Fri, May 29, 2020 at 06:09:06PM +0900, Masahiko Sawada wrote:
> > > Yes. Conversely, if we start logical replication in a physical
> > > replication connection (i.g. replication=true) we got a
Hi,
On 2020-06-03 19:28:12 +0200, Jerome Wagner wrote:
> I have been working on a node.js streaming client for different COPY
> scenarios.
> usually, during CopyOut, clients tend to buffer network chunks until they
> have gathered a full copyData message and pass that to the user.
>
> In some cas
On Thu, Jun 4, 2020 at 8:45 AM Andres Freund wrote:
> On 2020-06-03 14:19:45 -0400, Tom Lane wrote:
> > In the second place, this coding seems to me to indicate serious
> > confusion about which lock is protecting what. In the above example,
> > if writtenUpto is only accessed through atomic oper
Hello!
I'd like to propose a simple patch to allow for negative ISO 8601
intervals with leading minus, e.g. -PT1H besides PT-1H. It seems that
standard isn't quite clear on negative duration. However, lots of
software use leading minus and expect/generate intervals in such forms
making those incom
Hello,
I have been working on a node.js streaming client for different COPY
scenarios.
usually, during CopyOut, clients tend to buffer network chunks until they
have gathered a full copyData message and pass that to the user.
In some cases, this can lead to very large copyData messages. when ther
Hi,
On 2020-06-03 14:19:45 -0400, Tom Lane wrote:
> In connection with the nearby thread about spinlock coding rule
> violations, I noticed that we have several places that are doing
> things like this:
>
> SpinLockAcquire(&WalRcv->mutex);
> ...
> written_lsn = pg_atomic_read_u64
On Wed, Jun 3, 2020 at 03:07:30PM +0300, Ants Aasma wrote:
> On Tue, 2 Jun 2020 at 20:14, Bruce Momjian wrote:
>
> The server certificate should be issued by a certificate authority root
> outside of your organization only if you want people outside of your
> organization to trust yo
Hi,
On 2020-06-03 21:31:01 +0200, Tomas Vondra wrote:
> So there seems to be +40% between 9.6 and 10, and further +25% between
> 10 and master. However, plain hashagg, measured e.g. like this:
Ugh.
Since I am a likely culprit, I'm taking a look.
> Not sure what to think about this. Seems slot_
Hi,
When libpq is used to COPY data to the server, it doesn't properly
handle errors.
An easy way to trigger the problem is to start pgbench -i with a
sufficiently large scale, and then just shut the server down. pgbench
will happily use 100% of the cpu trying to send data to the server, even
tho
On Wed, Jun 3, 2020 at 09:16:03AM +0200, Fabien COELHO wrote:
> > > Also, I'm not at fully at ease with some of the underlying principles
> > > behind this proposal. Are we re-inventing/re-implementing kerberos or
> > > whatever? Are we re-implementing a brand new KMS inside pg? Why having
> > > o
Hi,
Not sure what's the root cause, but I can reproduce it. Timings for 9.6,
10 and master (all built from git with the same options) without explain
analyze look like this:
9.6
-
Time: 1971.314 ms
Time: 1995.875 ms
Time: 1997.408 ms
Time: 2069.913 ms
Time: 2004.196 ms
10
--
Hi,
On 2020-06-03 15:53:24 +0530, vignesh C wrote:
> Workers/
> Exec time (seconds) copy from file,
> 2 indexes on integer columns
> 1 index on text column copy from stdin,
> 2 indexes on integer columns
> 1 index on text column copy from file, 1 gist index on text column copy
> from file,
> 3 ind
Robert Haas writes:
> On Tue, Jun 2, 2020 at 5:04 AM Peter Eisentraut
> wrote:
>> The archeology reveals that these calls where originally added to
>> canonicalize the data_directory and config_file settings (7b0f060d54),
>> but that was then moved out of guc.c to be done early during postmaster
Hi,
On 2020-06-03 12:13:14 -0400, Robert Haas wrote:
> On Mon, May 18, 2020 at 12:48 AM Amit Kapila wrote:
> > In the above case, even though we are executing a single command from
> > the user perspective, but the currentCommandId will be four after the
> > command. One increment will be for th
Kyotaro Horiguchi writes:
> I looked through 224 locations where SpinLockAcquire and found some.
Yeah, I made a similar scan and arrived at about the same conclusions.
I think that the memcpy and strlcpy calls are fine; at least, we've got
to transport data somehow and it's not apparent why those
Jerome Wagner writes:
> now my question is the following :
> is it ok to consider that over the long term copyData is simply a transport
> container that exists only to allow the multiplexing of events in the
> protocol but that messages inside could be chunked over several copyData
> events ?
Ye
In connection with the nearby thread about spinlock coding rule
violations, I noticed that we have several places that are doing
things like this:
SpinLockAcquire(&WalRcv->mutex);
...
written_lsn = pg_atomic_read_u64(&WalRcv->writtenUpto);
...
SpinLockReleas
Hi,
On 2020-06-03 11:30:42 -0400, Robert Haas wrote:
> I too have seen recent benchmarking data where this was a big problem.
> Basically, you need a workload where the server doesn't have much or
> any actual query processing to do, but is just returning a lot of
> stuff to a really fast client -
Hackers,
The name "relkind" normally refers to a field of type 'char' with values like
'r' for "table" and 'i' for "index". In AlterTableStmt and CreateTableAsStmt,
this naming convention was abused for a field of type enum ObjectType. Often,
such fields are named "objtype", though also "kind
Hi Kyotaro-san,
Thank you for taking the time to review my patches. Would you like to
set yourself as a reviewer in the commit entry here?
https://commitfest.postgresql.org/28/2577/
> 0002:
>
> It is forgetting to grant pg_read_all_stats to execute
> pg_show_replication_origin_status. As the res
On Tue, Jun 2, 2020 at 5:04 AM Peter Eisentraut
wrote:
> The archeology reveals that these calls where originally added to
> canonicalize the data_directory and config_file settings (7b0f060d54),
> but that was then moved out of guc.c to be done early during postmaster
> startup (337ffcddba). The
On Mon, May 18, 2020 at 12:48 AM Amit Kapila wrote:
> In the above case, even though we are executing a single command from
> the user perspective, but the currentCommandId will be four after the
> command. One increment will be for the copy command and the other
> three increments are for lockin
st 3. 6. 2020 v 17:32 odesílatel Pavel Stehule
napsal:
> Hi
>
> One czech Postgres user reported performance issue related to speed
> HashAggregate in nested loop.
>
> The speed of 9.6
>
> HashAggregate (cost=27586.10..27728.66 rows=14256 width=24)
> (actual time=0.003..0.049 rows=39 loops=59920
On Wed, 3 Jun 2020 at 03:54, Michael Paquier wrote:
> Hi all,
>
> I have bumped into $subject, causing a replica identity index to
> be considered as dropped if running REINDEX CONCURRENTLY on it. This
> means that the old tuple information would get lost in this case, as
> a REPLICA IDENTITY US
Hi
One czech Postgres user reported performance issue related to speed
HashAggregate in nested loop.
The speed of 9.6
HashAggregate (cost=27586.10..27728.66 rows=14256 width=24)
(actual time=0.003..0.049 rows=39 loops=599203)
The speed of 10.7
HashAggregate (cost=27336.78..27552.78 rows=2160
On Tue, Jun 2, 2020 at 9:56 PM Andres Freund wrote:
> The biggest problem after that is that we waste a lot of time memcpying
> stuff around repeatedly. There is:
> 1) send function: datum -> per datum stringinfo
> 2) printtup: per datum stringinfo -> per row stringinfo
> 3) socket_putmessage: per
On Wed, 3 Jun 2020 at 03:14, Michael Paquier wrote:
> On Tue, Jun 02, 2020 at 04:46:55PM +0900, Masahiko Sawada wrote:
> > How about avoiding such an inconsistent situation? In that case,
> > replica identity works as NOTHING, but pg_class.relreplident is still
> > ‘i’, confusing users. It seems
On 2020/06/03 20:33, Dave Cramer wrote:
On Wed, 3 Jun 2020 at 01:19, Michael Paquier mailto:mich...@paquier.xyz>> wrote:
On Tue, Jun 02, 2020 at 02:23:50PM +0900, Fujii Masao wrote:
> On 2020/06/02 13:24, Michael Paquier wrote:
>> Still unconvinced as this restriction stands
Hi,
On 2020/06/03 11:14, Michael Paquier wrote:
Hi all,
I have been looking at the ODBC driver and the need for currtid() as
well as currtid2(), and as mentioned already in [1], matching with my
lookup of things, these are actually not needed by the driver as long
as we connect to a server ne
> On 3 Jun 2020, at 14:38, Michael Paquier wrote:
>
> On Mon, Jun 01, 2020 at 10:39:45AM +0200, Daniel Gustafsson wrote:
>> I don't have a problem with the existing wording of the first sentence, and I
>> don't have a problem with your suggestion either (as long as it's parameters
>> in
>> plura
On Mon, Jun 01, 2020 at 10:39:45AM +0200, Daniel Gustafsson wrote:
> I don't have a problem with the existing wording of the first sentence, and I
> don't have a problem with your suggestion either (as long as it's parameters
> in
> plural).
Thanks, that's why I kept the word plural in my own sug
Hi hackers,
Partitioning is necessary for very large tables.
However, I found that postgresql does not support create index concurrently on
partitioned tables.
The document show that we need to create an index on each partition
individually and then finally create the partitioned index non-conc
On Tue, 2 Jun 2020 at 20:14, Bruce Momjian wrote:
> The server certificate should be issued by a certificate authority root
> outside of your organization only if you want people outside of your
> organization to trust your server certificate, but you are then asking
> for the client to only trus
On Wed, 3 Jun 2020 at 01:19, Michael Paquier wrote:
> On Tue, Jun 02, 2020 at 02:23:50PM +0900, Fujii Masao wrote:
> > On 2020/06/02 13:24, Michael Paquier wrote:
> >> Still unconvinced as this restriction stands for logical decoding
> >> requiring a database connection but it is not necessarily
On Fri, May 29, 2020 at 8:31 PM Dilip Kumar wrote:
>
The fixes in the latest patchset are correct. Few minor comments:
v26-0005-Implement-streaming-mode-in-ReorderBuffer
+ /*
+ * Mark toplevel transaction as having catalog changes too if one of its
+ * children has so that the ReorderBufferBuild
On Wed, 3 Jun 2020 at 16:16, Fabien COELHO wrote:
>
>
> Hello Masahiko-san,
>
> > This key manager is aimed to manage cryptographic keys used for
> > transparent data encryption. As a result of the discussion, we
> > concluded it's safer to use multiple keys to encrypt database data
> > rather tha
On 2020/06/03 12:06, Kyotaro Horiguchi wrote:
At Wed, 3 Jun 2020 09:43:17 +0900, Fujii Masao
wrote in
I will change the status back to Needs Review.
Thanks for the review!
record = ReadCheckpointRecord(xlogreader, checkPointLoc, 1, false);
if (record != NULL)
On 3/30/20 1:15 PM, Etsuro Fujita wrote:
Hi,
On Wed, Mar 11, 2020 at 10:47 AM movead li wrote:
I redo the make installcheck-world as Kyotaro Horiguchi point out and the
result nothing wrong. And I think the patch is good in feature and performance
here is the test result thread I made before:
On Tue, Jun 2, 2020 at 7:53 PM Dilip Kumar wrote:
>
> On Tue, Jun 2, 2020 at 4:56 PM Amit Kapila wrote:
> >
> > On Tue, Jun 2, 2020 at 3:59 PM Dilip Kumar wrote:
> > >
> > > I thin for our use case BufFileCreateShared is more suitable. I think
> > > we need to do some modifications so that we c
I got it should be LSN + MAXALIGN(xlogrecord length) 👍
Thanks a lot.
> On 2 Jun 2020, at 19:11, Jehan-Guillaume de Rorthais wrote:
>
> Nope, just sum the xlogrecord length.
At Tue, 2 Jun 2020 13:13:18 -0300, Martín Marqués
wrote in
> > > I have no problem adding it to this ROLE, but we'd have to amend the
> > > doc for default-roles to reflect that SELECT for this view is also
> > > granted to `pg_read_all_stats`.
> >
> > I agree in general that pg_monitor shouldn'
Hello Masahiko-san,
This key manager is aimed to manage cryptographic keys used for
transparent data encryption. As a result of the discussion, we
concluded it's safer to use multiple keys to encrypt database data
rather than using one key to encrypt the whole thing, for example, in
order to ma
On Thu, May 28, 2020 at 11:08 PM Ashutosh Bapat
wrote:
> On Wed, May 27, 2020 at 6:51 PM Amit Langote wrote:
> > So in Rajkumar's example, the cursor is declared as:
> >
> > CURSOR IS SELECT * FROM tbl WHERE c1< 5 FOR UPDATE;
> >
> > and the WHERE CURRENT OF query is this:
> >
> > UPDATE tbl SET
75 matches
Mail list logo
