Peter Geoghegan writes:
> On Tue, Sep 6, 2016 at 6:17 AM, Tom Lane wrote:
>> It doesn't seem to me that this limit has anything to do with anything,
>> and the comment claiming only that it's "noncritical" isn't helping.
> You've not understood the problem
On 9/5/16 12:14 PM, Bruce Momjian wrote:
> I have certainly faced my fair share of customers with dangling temp
> tables, and would like to see this changed in some way or another.
I don't think we look at those temp tables frequently enough to justify
keeping them around for all users.
Plus,
On 8/20/16 3:05 PM, Tom Lane wrote:
> Make initdb's suggested "pg_ctl start" command line more reliable.
>
> The original coding here was not nearly careful enough about quoting
> special characters, and it didn't get corner cases right for constructing
> the pg_ctl path either. Use
On 9/4/16 5:10 PM, Christian Convey wrote:
Thanks for the response. I'm unclear about how the TODO list is
curated. Is there someone whose attention I should direct to this
thread?
There's a number of people that can edit the TODO list. I've removed
that item.
--
Jim Nasby, Data Architect,
On 9/4/16 1:46 AM, Simon Riggs wrote:
> The patch also makes vacuum free the dead_tuples before starting
> truncation. It didn't seem necessary to hold onto it beyond that
> point, and it might help give the OS more cache, especially if work
> mem is configured very high to avoid multiple index
On 9/3/16 1:30 PM, Tom Lane wrote:
Or we could add something like "But first, we must deal with the special
case of reversing direction after reaching EOF."
I'm working on that, but one thing isn't clear to me... why do we only
skip past the last tuple if (!node->leader->eof_cte)? Even if
On 8/11/16 9:12 PM, Michael Paquier wrote:
> Note that pg_dump[all] and pg_upgrade already have safeguards against
> those things per the same routines putting quotes for execution as
> commands into psql and shell. So attached is a patch to implement this
> restriction in the backend, and I am
Hello list
While testing an application with PostgreSQL 9.5, we experienced an issue
involving aborted subtransactions and SELECT FOR UPDATE. In certain
situations, a locking query doesn't return rows that should be visible and
already locked by the current transaction.
I bisected this down to
On Tue, Sep 6, 2016 at 11:08 PM, Simon Riggs wrote:
> On 29 August 2016 at 14:52, Fujii Masao wrote:
>> On Sat, Aug 6, 2016 at 6:36 PM, Petr Jelinek wrote:
>>> On 04/08/16 06:40, Masahiko Sawada wrote:
On Wed, Aug 3,
On 8/31/16 1:32 PM, Peter Geoghegan wrote:
> ICU is happy to have multiple versions
> of a collation at a time, and you'll probably retain the old collation
> version in ICU.
>
> Even if your old collation version isn't available in a new ICU
> release (which I think is unlikely in practice)
I
On Tue, Sep 6, 2016 at 6:17 AM, Tom Lane wrote:
> It doesn't seem to me that this limit has anything to do with anything,
> and the comment claiming only that it's "noncritical" isn't helping.
> If the point is to prevent the later LACKMEM check from failing, then
> certainly
On 8/31/16 4:24 PM, Doug Doole wrote:
> ICU explicitly does not provide stability in their locales and collations. We
> pushed them hard to provide this, but between changes to the CLDR data and
> changes to the ICU code it just wasn’t feasible for them to provide version
> to version
On 8/31/16 12:32 AM, Michael Paquier wrote:
> On Wed, Aug 31, 2016 at 1:12 PM, Peter Eisentraut
> wrote:
>> On 8/30/16 11:27 PM, Craig Ringer wrote:
>>> Speaking of which, have you had a chance to try it on Windows yet?
>>
>> nope
>
> +SELECT a, b FROM
On 9/1/16 9:53 AM, Peter Eisentraut wrote:
On 8/15/16 3:39 PM, David Steele wrote:
That patch got me thinking about what else could be excluded and after
some investigation I found the following: pg_notify, pg_serial,
pg_snapshots, pg_subtrans. These directories are all cleaned, zeroed,
or
06.09.2016 07:44, Pavan Deolasee:
On Mon, Aug 8, 2016 at 3:13 PM, Anastasia Lubennikova
>
wrote:
Thank you for the review, I'll fix these problems in final version.
Posting the first message I intended to raise
Hi,
Changes look good to me.
However there are couple of minor issues need to be fixed.
1.
"under" repeated on second line. Please remove.
+if and when CustomScanState is located under
+under LimitState; which implies the underlying node is
not
2.
Typo: dicsussion => discussion
Please
On 6 September 2016 at 11:30, Simon Riggs wrote:
> In vacuumlazy.c, VACUUM_TRUNCATE_LOCK_WAIT_INTERVAL is described as
> being in ms on line 85, yet it is used on line 1759 in a call to
> pg_usleep, so is treated as microseconds rather than milliseconds.
>
> As a result,
Hello,
I was able to find a patch file where there is a call to ExitPostmaster() in
postmaster.c .
@@ -3081,6 +3081,11 @@
shmem_exit(1);
reset_shared(PostPortNumber);
+ /* recovery termination */
+ ereport(FATAL,
+ (errmsg("recovery termination due to
On 9/6/16 8:07 AM, Robert Haas wrote:
On Wed, Aug 31, 2016 at 9:45 PM, Simon Riggs wrote:
>
Related cleanup
* Promotion signal file is now called "promote.trigger" rather than
just "promote"
* Remove user configurable "trigger_file" mechanism - use
"promote.trigger"
On 29 August 2016 at 14:52, Fujii Masao wrote:
> On Sat, Aug 6, 2016 at 6:36 PM, Petr Jelinek wrote:
>> On 04/08/16 06:40, Masahiko Sawada wrote:
>>>
>>> On Wed, Aug 3, 2016 at 3:05 PM, Michael Paquier
>>> wrote:
On Tue, Sep 6, 2016 at 10:01 PM, Petr Jelinek wrote:
> That's also reasonable solution, I don't really have preference between
> those. My main point was to get rid of the 5 or so variables where only one
> will actually be used in the end.
And no need to worry about the
Peter Geoghegan writes:
> While working on my parallel CREATE INDEX patch, I came across a
> problem.
I took a quick look at this. I do not follow the logic in this new bit:
+ if (availMemLessRefund <=
+ (int64) state->activeTapes *
On Tue, Sep 6, 2016 at 5:29 PM, Robert Haas wrote:
> On Mon, Sep 5, 2016 at 3:17 PM, Amit Kapila wrote:
>> On Mon, Aug 15, 2016 at 10:35 PM, Robert Haas wrote:
> Don't you need to set proc->cvSleeping = false in
On Tue, Sep 6, 2016 at 9:19 PM, Robert Haas wrote:
> On Wed, Aug 31, 2016 at 1:05 PM, Amit Langote
> wrote:
>>> However, it seems a lot better to make it a property of the parent
>>> from a performance point of view. Suppose there are 1000
On Fri, Sep 2, 2016 at 3:55 PM, Etsuro Fujita
wrote:
> Hi Ashutosh,
>
> On 2016/08/22 15:49, Ashutosh Bapat wrote:
>
>> 1. deparsePlaceHolderVar looks odd - each of the deparse* function is
>> named as deparse + > into>. PlaceHolderVar is not a parser node, so no
Hi,
On Tue, Sep 6, 2016 at 8:15 PM, Rajkumar Raghuwanshi
wrote:
> Hi,
>
> I have applied updated patches given by you, and observe below.
>
> here in the given example, t6_p3 partition is not allowed to have null, but
> I am able to insert it, causing two
On 06/09/16 13:52, David Steele wrote:
On 9/6/16 4:56 AM, Petr Jelinek wrote:
On 06/09/16 07:18, Abhijit Menon-Sen wrote:
Do we want something like this (easy to implement and document, perhaps
not especially convenient to use):
recovery_target = 'xid' # or
Hi,
Attached is a WIP patch extending the postgres_fdw DML pushdown in 9.6
so that it can perform an update/delete on a join remotely. An example
is shown below:
* without the patch:
postgres=# explain verbose delete from ft1 using ft2 where ft1.a = ft2.a;
QUERY PLAN
(Just remembered to remove pgsql-committers here).
On Tue, Sep 6, 2016 at 9:21 PM, Christian Ullrich wrote:
> * Michael Paquier wrote:
>
>> On Tue, Sep 6, 2016 at 5:36 PM, Christian Ullrich
>> wrote:
>>>
>>> * Michael Paquier wrote:
>
In order to
This patch uses RangeBound structure. There's also a structure defined with
the same name in rangetypes.h with some slight differences. Should we
rename the one in partition.c as PartRangeBound or something like that to
avoid the confusion?
--
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
* Michael Paquier wrote:
On Tue, Sep 6, 2016 at 5:36 PM, Christian Ullrich wrote:
* Michael Paquier wrote:
In order to avoid any problems with the load and unload windows, my
bet goes for 0001, 0002 and 0003, with the last two patches merged
together, 0001 being only
While reviewing Peter's latest round of sorting patches, and trying to
understand the new "batch allocation" mechanism, I started to wonder how
useful the pre-reading in the merge stage is in the first place.
I'm talking about the code that reads a bunch of from each tape, loading
them into
On Wed, Aug 31, 2016 at 1:05 PM, Amit Langote
wrote:
>> However, it seems a lot better to make it a property of the parent
>> from a performance point of view. Suppose there are 1000 partitions.
>> Reading one toasted value for pg_class and running stringToNode()
On Tue, Sep 6, 2016 at 4:00 PM, Simon Riggs wrote:
> I'm also wondering why we don't use lock_timeout when the user sets it?
> Not a bug, but patch attached anyway.
> vacuum_truncate_use_lock_timeout.v1.patch
This part seems fairly random. I don't think it makes sense to
On Wed, Aug 31, 2016 at 9:45 PM, Simon Riggs wrote:
> This is a summary of proposed changes to the recovery.conf API for
> v10. These are based in part on earlier discussions, and represent a
> minimal modification to current usage.
This looks great.
> Proposed changes
On Tue, Sep 6, 2016 at 5:36 PM, Christian Ullrich wrote:
> * Michael Paquier wrote:
>> In order to avoid any problems with the load and unload windows, my
>> bet goes for 0001, 0002 and 0003, with the last two patches merged
>> together, 0001 being only a set of independent
On Mon, Sep 5, 2016 at 3:17 PM, Amit Kapila wrote:
> On Mon, Aug 15, 2016 at 10:35 PM, Robert Haas wrote:
Don't you need to set proc->cvSleeping = false in ConditionVariableSignal?
>>>
>>> I poked at this a bit... OK, a lot... and have some
On 9/6/16 4:56 AM, Petr Jelinek wrote:
On 06/09/16 07:18, Abhijit Menon-Sen wrote:
Do we want something like this (easy to implement and document, perhaps
not especially convenient to use):
recovery_target = 'xid' # or 'time'/'name'/'lsn'/'immediate'
recovery_target_xid = xxx? #
"K S, Sandhya (Nokia - IN/Bangalore)" writes:
> I was able to find a patch file where there is a call to ExitPostmaster() in
> postmaster.c .
> @@ -3081,6 +3081,11 @@
> shmem_exit(1);
> reset_shared(PostPortNumber);
> + /* recovery termination */
Hi,
I have applied updated patches given by you, and observe below.
here in the given example, t6_p3 partition is not allowed to have null, but
I am able to insert it, causing two nulls in the table.
--create a partition table
create table t6 (a int, b varchar) partition by list(a);
create
On 6 September 2016 at 16:10, Daniel Verite wrote:
> Craig Ringer wrote:
>
>> Updated patch attached.
>
> Please find attached a couple fixes for typos I've came across in
> the doc part.
Thanks, will apply and post a rebased patch soon, or if someone picks
this
In vacuumlazy.c, VACUUM_TRUNCATE_LOCK_WAIT_INTERVAL is described as
being in ms on line 85, yet it is used on line 1759 in a call to
pg_usleep, so is treated as microseconds rather than milliseconds.
As a result, the timeout during lazy_truncate_heap() is actually only
5ms long, not 5s long.
So
On 6 Sep. 2016 17:57, "Pavan Deolasee" wrote:
>
>
>
> On Tue, Sep 6, 2016 at 3:06 PM, Craig Ringer
wrote:
>>
>>
>> I think it's worth looking at how Java handles logging. We can't achieve
an exact parallel in C as we don't really have a
On 1 September 2016 at 21:28, Simon Riggs wrote:
> So the only way to handle multiple locks is to do this roughly the way
> Rod suggests.
I've just been reading the VACUUM code and it turns out that we
already use Rod's mechanism internally. So on that basis it seems fine
On Tue, Sep 6, 2016 at 3:06 PM, Craig Ringer
wrote:
>
> I think it's worth looking at how Java handles logging. We can't achieve
> an exact parallel in C as we don't really have a class hierarchy ... but we
> do have subsystems roughly grouped by file and directory
On 05/09/16 23:35, Steve Singer wrote:
On 09/05/2016 03:58 PM, Steve Singer wrote:
On 08/31/2016 04:51 PM, Petr Jelinek wrote:
Hi,
and one more version with bug fixes, improved code docs and couple
more tests, some general cleanup and also rebased on current master
for the start of CF.
> On 06 Sep 2016, at 12:09, Simon Riggs wrote:
>
> On 6 September 2016 at 09:58, Stas Kelvich wrote:
>>
>> I'll check it against my failure scenario with subtransactions and post
>> results or updated patch here.
>
> Make sure tests are added
On 05/09/16 21:58, Steve Singer wrote:
On 08/31/2016 04:51 PM, Petr Jelinek wrote:
Hi,
and one more version with bug fixes, improved code docs and couple
more tests, some general cleanup and also rebased on current master
for the start of CF.
To get the 'subscription' TAP tests to pass I
On 6 Sep. 2016 17:28, "Pavan Deolasee" wrote:
>
>
> The patch uses some preprocessing and scripting magic to assign distinct
identifiers to each module (a subdir in the source code), to each file and
to each elog message. It then provides a set of functions by which an
While reviewing Jeff's notice_lock_waits patch, I came across his comment
about having a general facility for promoting selected LOG messages. So I
thought I should post it here, even though the patch is probably far from
being accepted in Postgres.
I recently wrote a patch for Postgres-XL to do
At 2016-09-06 10:56:53 +0200, p...@2ndquadrant.com wrote:
>
> So +1 on the recovery_target = 'xid:xxx' idea.
OK, I'll make it work that way. New patch in a couple of days.
-- Abhijit
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
On 6 September 2016 at 09:58, Stas Kelvich wrote:
>> On 06 Sep 2016, at 04:41, Michael Paquier wrote:
>>
>> On Sat, Sep 3, 2016 at 10:26 PM, Michael Paquier
>> wrote:
>>> On Fri, Sep 2, 2016 at 5:06 AM, Simon Riggs
On Tue, Sep 6, 2016 at 5:58 PM, Stas Kelvich wrote:
>> On 06 Sep 2016, at 04:41, Michael Paquier wrote:
>> This overlaps with TwoPhaseGetGXact but I'd rather keep both things
>> separated: it does not seem worth complicating the existing
On 6 Sep. 2016 15:12, "Michael Paquier" wrote:
>
> On Thu, Sep 1, 2016 at 4:03 PM, Christian Ullrich
wrote:
>
> > That said, introducing this requirement would be a very significant
change.
> > I'm not sure how many independently maintained
> On 06 Sep 2016, at 04:41, Michael Paquier wrote:
>
> On Sat, Sep 3, 2016 at 10:26 PM, Michael Paquier
> wrote:
>> On Fri, Sep 2, 2016 at 5:06 AM, Simon Riggs wrote:
>>> On 13 April 2016 at 15:31, Stas Kelvich
On 06/09/16 07:18, Abhijit Menon-Sen wrote:
One open issue is the various assign_recovery_target_xxx functions,
which Michael noted in his earlier review:
+static void
+assign_recovery_target_xid(const char *newval, void *extra)
+{
+ recovery_target_xid = *((TransactionId *) extra);
+
+
* Michael Paquier wrote:
On Thu, Sep 1, 2016 at 4:03 PM, Christian Ullrich wrote:
My conclusion from April stands:
The fact that master looks like it does means that there have not been
many (or any) complaints about missing cross-module environment
variables. If
Hello, Christian
> Can anyone suggest a project for my first PG contribution?
In my experience good starting points are:
* Doing code review. It's really important for PostgreSQL project.
Everyone likes to write code, no one likes to read or test it :)
* Finding typos, there are a lot of
On 6 September 2016 at 08:06, Abhijit Menon-Sen wrote:
> At 2016-09-06 14:40:54 +0900, michael.paqu...@gmail.com wrote:
>>
>> my best advice here is to make all those recovery_target_* parameters
>> PGC_POSTMASTER so as they are loaded only once when the server starts,
>>
Craig Ringer wrote:
> Updated patch attached.
Please find attached a couple fixes for typos I've came across in
the doc part.
Also it appears that PQqueriesInBatch() doesn't work as documented.
It says:
"Returns the number of queries still in the queue for this batch"
but in fact it's
On Fri, Aug 5, 2016 at 10:30 PM, Jeff Janes wrote:
>
>
> A general facility for promoting selected LOG messages to NOTICE would
> be nice, but I don't know how to design or implement that. This is
> much easier, and I find it quite useful.
>
>
IMHO that's what we need and
On 08/31/2016 05:54 PM, Craig Ringer wrote:
How do you get the commit LSN to watch for? Grab
pg_current_xlog_insert_location() just after the commit and figure
that replaying to that point guarantees you get the commit?
That's the point, it was created in order to provide the cosistent view
I'm reviewing patches 1-3 in this series, i.e. those patches that are
not directly related to parallelism, but are independent improvements to
merging.
Let's begin with patch 1:
On 08/02/2016 01:18 AM, Peter Geoghegan wrote:
Cap the number of tapes used by external sorts
Commit df700e6b set
I found a server crash while running make check in regress folder. with
this set of patches. Problem is RelationBuildPartitionKey() partexprsrc may
be used uninitialized. Initializing it with NIL fixes the crash. Here's
patch to fix it. Came up with the fix after discussion with Amit.
On Fri, Aug
On Thu, Sep 1, 2016 at 4:03 PM, Christian Ullrich wrote:
> * Michael Paquier wrote:
>
>> On Wed, Apr 27, 2016 at 2:39 AM, Christian Ullrich
>> wrote:
>
>>> * Christian Ullrich wrote:
>
>> And actually, by looking at those patches, isn't it a dangerous
On 08/16/2016 03:33 AM, Peter Geoghegan wrote:
I attach a patch that changes how we maintain the heap invariant
during tuplesort merging. I already mentioned this over on the
"Parallel tuplesort, partitioning, merging, and the future" thread. As
noted already on that thread, this patch makes
At 2016-09-06 14:40:54 +0900, michael.paqu...@gmail.com wrote:
>
> my best advice here is to make all those recovery_target_* parameters
> PGC_POSTMASTER so as they are loaded only once when the server starts,
> and then we define the recovery target type used in the startup
> process instead of
On 6 September 2016 at 02:41, Michael Paquier wrote:
> After review the result is attached. Perhaps a committer could get a look at
> it?
Yes, will do, but it will be a few more days yet.
--
Simon Riggshttp://www.2ndQuadrant.com/
PostgreSQL
On 5 September 2016 at 21:58, Claudio Freire wrote:
How long does that part ever take? Is there any substantial gain from this?
> Btw, without a further patch to prefetch pages on the backward scan
> for truncate, however, my patience ran out before it finished
>
101 - 169 of 169 matches
Mail list logo