Thanks for the comments.
I'll revise the patch along the discussion. Before starting code work,
please let me summarize the discussion.
* Generally, we should keep FDWs away from PostgreSQL internals,
such as TupleTableSlot.
* FDW should have planner hook which allows FDW to create FDW-specifi
Thank you, but I think about this last night. Opening unnecessary portals
isn't good idea, similarly sending 2nd describe when statement was
prepared. Currently JDBC drivers doesn't make this. I think better will be
to store what format we had requested on stack, and then coerce those
formats when
Simon Riggs writes:
> That would mean running GetCurrentTransactionId() inside LockAcquire()
> if (lockmode >= AccessExclusiveLock &&
> locktag->locktag_type == LOCKTAG_RELATION &&
> !RecoveryInProgress())
> (void) GetCurrentTransactionId();
> Any objections to that fix?
Could we
Maciek Sakrejda writes:
>> 21:43:02.264 (26) FE=> Describe(statement=S_1)
> You're still doing the statement-flavor Describe. As Tom pointed out,
> this won't tell you the result types because it doesn't know them.
> Actually, technically if you issue a statement-flavor Describe *after*
> a Bind,
On Fri, Nov 26, 2010 at 3:11 AM, Bruce Momjian wrote:
> I have applied this patch, with modified wording of the "cannot connect"
> case:
>
> $ pg_ctl -w -l /dev/null start
> waiting for server to start done
> server started
> warning: could not connect, perhaps due
On Thu, 25 Nov 2010 18:40:09 -0800
David Fetter wrote:
> On Thu, Nov 25, 2010 at 05:51:11PM +0900, Shigeru HANADA wrote:
> > I'm going to add new CommitFest items for this patch and "SQL/MED -
> > postgresql_fdw" patch which have been split from "SQL/MED" patch. Can
> > I add them to CF 2010-11 w
(2010/11/26 11:35), Fujii Masao wrote:
On Thu, Nov 25, 2010 at 3:18 PM, KaiGai Kohei wrote:
The attached patch is revised version.
- Logging part within auth_delay was removed. This module now focuses on
injection of a few seconds delay on authentication failed.
- Documentation parts were ad
On Thu, Nov 25, 2010 at 05:51:11PM +0900, Shigeru HANADA wrote:
> On Thu, 25 Nov 2010 17:12:44 +0900
> Shigeru HANADA wrote:
> > Attached is a patch that adds file_fdw, FDW which reads records from
> > files on the server side, as a contrib module. This patch is based on
> > "SQL/MED core functio
On Thu, Nov 25, 2010 at 3:18 PM, KaiGai Kohei wrote:
> The attached patch is revised version.
>
> - Logging part within auth_delay was removed. This module now focuses on
> injection of a few seconds delay on authentication failed.
> - Documentation parts were added like any other contrib modules
On Thu, Nov 25, 2010 at 10:24:51PM +0100, Dimitri Fontaine wrote:
> Joshua Tolley writes:
> > I've just looked at pg_execute_from_file[1]. The idea here is to execute all
> > the SQL commands in a given file. My comments:
>
> Thanks for your review. Please find attached a revised patch where I've
On Fri, Nov 26, 2010 at 7:40 AM, Simon Riggs wrote:
> As to solutions, it cannot be acceptable to ignore some locks just
> because an xid has not been assigned.
Even if GetRunningTransactionLocks ignores such a lock, it's eventually
WAL-logged by LogAccessExclusiveLock, isn't it?
> If I understa
(2010/10/16 4:49), Josh Kupershmidt wrote:
> [Moving to -hackers]
>
> On Fri, Oct 15, 2010 at 3:43 AM, Simon Riggs wrote:
>> On Mon, 2010-10-11 at 09:41 -0400, Josh Kupershmidt wrote:
>>> On Thu, Oct 7, 2010 at 7:43 PM, Josh Kupershmidt wrote:
>>>
I noticed that granting a user column-level
Hi,Robert.
Thanks for your reply.
As far as I looked into postgre's source,
I came to know that It seems that it uses strcoll to compare strings.
So it depends on the underlying operating system,like you said.
Charles.
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To ma
On Thu, 2010-11-25 at 16:59 +0900, Fujii Masao wrote:
> To solve the problem, ISTM that XID should be assigned before the
> information about AccessExclusive lock becomes visible to another
> process. Or CHECKPOINT (i.e., GetRunningTransactionLocks) should
> ignore the locks with XID = 0.
First,
Robert Haas writes:
> One thing I've often wished for is the ability to dump a specific
> function
See getddl from OmniTI, or the alternative version I kept forgetting to
put online somewhere:
https://labs.omniti.com/labs/pgtreats/wiki/getddl
https://github.com/dimitri/getddl
The OmniTI ver
Hi,
So I've been working on improving locks for foreign key checks, as
discussed in a thread started by Joel Jacobson a while ago. I've posted
about this:
http://www.commandprompt.com/blogs/alvaro_herrera/2010/11/fixing_foreign_key_deadlocks/
http://www.commandprompt.com/blogs/alvaro_herrera/2010
Robert Haas writes:
> Please do. Tab completion support should really be included in the
> patch - adding it as a separate patch is better than not having it, of
> course.
Please find attached version 9 of the patch, which includes psql
completion support of the "SET SCHEMA" variant of already s
Joshua Tolley writes:
> I've just looked at pg_execute_from_file[1]. The idea here is to execute all
> the SQL commands in a given file. My comments:
Thanks for your review. Please find attached a revised patch where I've
changed the internals of the function so that it's split in two and that
th
Itagaki Takahiro writes:
> Thanks. I'll move the patch to Ready for Committer.
Thanks!
--
Dimitri Fontaine
http://2ndQuadrant.fr PostgreSQL : Expertise, Formation et Support
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://w
> 21:43:02.264 (26) FE=> Describe(statement=S_1)
You're still doing the statement-flavor Describe. As Tom pointed out,
this won't tell you the result types because it doesn't know them.
Actually, technically if you issue a statement-flavor Describe *after*
a Bind, the server does have this inform
On 10-11-22 03:24 PM, Steve Singer wrote:
On 10-11-22 09:37 AM, Gurjeet Singh wrote:
On Sat, Nov 20, 2010 at 9:00 AM, Steve Singer
Almost fixed.
I still get an unexpected difference.
! DETAIL: cannot create PRIMARY KEY/UNIQUE constraint with a non-unique
index.
CREATE UNIQUE INDEX rpi_idx2 O
Hm... I moved Bind before Describe, I now have
// Construct a new portal if needed.
Portal portal = null;
if (usePortal)
{
String portalName = "C_" + (nextUniqueID++);
portal = new Portal(query, portalName);
}
sendBind(query,
> OTOH, it seems possible that the JDBC driver might behave differently
> depending on whether parameter types were prespecified or not --- it
> might issue Describe earlier in order to get the parameter types,
> perhaps.
Ah. Bingo:
boolean describeStatement = describeOnly || (!oneShot &&
Maciek Sakrejda writes:
>> But to the last part of cited protocol specification, when I've sent message
>> with statement parameter's type int4, int8, varchar the format
>> field wasn't set to 0, but 1.
> I wasn't able to reproduce that with my standalone test case. When I
> changed the paramete
On Thu, 25 Nov 2010 11:28:02 -0800, Maciek Sakrejda
wrote:
>> So, to summarise, I shouldn't believe server DescribeRow (in context of
>> format), in this situation, but only I should look at this what I asked
>> for, isn't it? If I asked for columns in binary format, I need to do
>> binary reading
> So, to summarise, I shouldn't believe server DescribeRow (in context of
> format), in this situation, but only I should look at this what I asked
> for, isn't it? If I asked for columns in binary format, I need to do binary
> reading regarding what server has responded?
Yes, because in this ca
On Thu, Nov 25, 2010 at 1:40 AM, Chang Chao wrote:
> How strings are sorted when LC_COLLATE = ja_JP.UTF-8.
> I tried to read the documention on that,but there are just a few words,
> like LC_COLLATE determines string sort order,
> Is there a specific reference about this?
> So I can implement an e
On Thu, Nov 25, 2010 at 11:23 AM, Tom Lane wrote:
> Robert Haas writes:
>> No, what I was suggesting was taking the existing function:
>> extern void pgstat_report_waiting(bool waiting);
>> ...and instead doing something like this:
>> extern void pgstat_report_waiting(char *reason);
>> ...and the
Hi,
Thank you for your response.
I would only ask to be sure...
So, to summarise, I shouldn't believe server DescribeRow (in context of
format), in this situation, but only I should look at this what I asked
for, isn't it? If I asked for columns in binary format, I need to do binary
reading regar
Okay, looking at the JDBC side of things, I think JDBC doesn't
actually need that information (since, it always used text results
before Radosław's patch--the previous binary support was for
parameters only, right?). From looking at QueryExecutorImpl
(specifically sendOneQuery), it's clear that it
Bruce Momjian wrote:
> > > BTW, it is annoying that we can't definitively distinguish "postmaster
> > > is not running" from a connectivity problem, but I can't see a way
> > > around that.
> >
> > Agreed. I will research this.
>
> I have researched this and developed the attached patch. It imp
Bruce Momjian wrote:
> Win32 buildfarm members are red because of my inet_pton changes. I will
> look into this in the next day, and also improve how we include C files
> from /port for libpq.
OK, I have accomplished both goals with the two attached, applied
patches.
--
Bruce Momjian
Tom Lane writes:
> Just make two pg_proc entries that are pointing at two C functions.
> The C functions can call a common subroutine after extracting their
> arguments.
Mmmm, ok, will adapt the idea to the current code, where the extracting
is mingled into the processing. Thanks for the idea, th
Interesting. I think you're right. Looking at the Wireshark traffic
again, the driver seems to issue a portal-variant Describe when using
unnamed prepared statements, but as soon as the named prepared
statements kick in (per prepare threshold), the Describe is a
statement-variant Describe with the
Dimitri Fontaine writes:
> The pg_execute_from_file() function is designed to work with either a
> filename as its sole argument, or the filename and a VARIADIC text list
> of arguments containing placeholder names and values. It works fine with
> two entries in pg_proc using the same backend func
Heikki Linnakangas writes:
> On 25.11.2010 18:28, Tom Lane wrote:
>> Or just specify a format for the extra information. Perhaps it could be
>> thought of as being a value of type bytea? Obviously we can't just have
>> a fixed amount of info, but maybe a blob with a length word is enough.
> Tha
Maciek Sakrejda writes:
> Since triggering the set of FEBE messages that leads to this was tied
> deep into the guts of JDBC, I opted for raw wire protocol. It looks
> like the following sequence of messages from the client leads to this
> result format mixup:
> 1. Parse, with statement name "S_1
Hi,
Trying to fix a regression test problem I've left for better days while
developping the extensions, some help is needed.
The pg_execute_from_file() function is designed to work with either a
filename as its sole argument, or the filename and a VARIADIC text list
of arguments containing placeh
2010/11/25 KaiGai Kohei :
> The attached patch is a revised patch.
>
> - The utils/hooks.h was renamed to catalog/objectaccess.h
> - Numeric in the tail of InvokeObjectAccessHook0() has gone.
> - Fixed bug in ATExecAddColumn; it gave AttributeRelationId
> to the hook instead of RelationRelationId.
On 25.11.2010 18:28, Tom Lane wrote:
Heikki Linnakangas writes:
Hmm, I see, cached plans are planned in a shorter-lived context first,
and copied to permanent storage afterwards. Needs more thought then.
Maybe the FDW needs to provide a copyFdwPlan() function to copy FdwPlans
returned by that F
Excerpts from Tom Lane's message of jue nov 25 13:23:42 -0300 2010:
> Robert Haas writes:
> > No, what I was suggesting was taking the existing function:
> > extern void pgstat_report_waiting(bool waiting);
> > ...and instead doing something like this:
> > extern void pgstat_report_waiting(char *r
Heikki Linnakangas writes:
> Hmm, I see, cached plans are planned in a shorter-lived context first,
> and copied to permanent storage afterwards. Needs more thought then.
> Maybe the FDW needs to provide a copyFdwPlan() function to copy FdwPlans
> returned by that FDW.
Or just specify a format
On 25.11.2010 18:18, Tom Lane wrote:
Heikki Linnakangas writes:
I left out some details on what exactly FdwPlan should contain and what
it's lifecycle should be. I'm thinking that it should be allocated in
the CurrentMemoryContext that's active when the FDW Plan routine is
called, which would b
Robert Haas writes:
> No, what I was suggesting was taking the existing function:
> extern void pgstat_report_waiting(bool waiting);
> ...and instead doing something like this:
> extern void pgstat_report_waiting(char *reason);
> ...and then arrange to pass the reason via the eponymous argument.
Heikki Linnakangas writes:
> I left out some details on what exactly FdwPlan should contain and what
> it's lifecycle should be. I'm thinking that it should be allocated in
> the CurrentMemoryContext that's active when the FDW Plan routine is
> called, which would be the same context where we s
Alvaro Herrera writes:
> On the other hand, pg_locks is already rather unwieldy to use. We
> already have a self-join that tells us the details of what's locking
> processes: you need to join pg_locks like this:
> ...
> and throw in a bunch of left joins to see the details of database,
> relation
Win32 buildfarm members are red because of my inet_pton changes. I will
look into this in the next day, and also improve how we include C files
from /port for libpq.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossi
On Thu, Nov 25, 2010 at 10:05 AM, Alvaro Herrera
wrote:
> Excerpts from Robert Haas's message of jue nov 25 11:56:27 -0300 2010:
>
>> No, what I was suggesting was taking the existing function:
>>
>> extern void pgstat_report_waiting(bool waiting);
>>
>> ...and instead doing something like this:
>
On Thu, Nov 25, 2010 at 4:45 AM, Koichi Suzuki wrote:
>>> plus the
>>> communication protocol overhead and latency. However, it occurs to me
>>> that if you had a shared disk system via either iSCSI, Fiber Channel,
>>> NFS, or whatever (which also had higher I/O capabilities than a single
>>> serv
Excerpts from Robert Haas's message of jue nov 25 11:56:27 -0300 2010:
> No, what I was suggesting was taking the existing function:
>
> extern void pgstat_report_waiting(bool waiting);
>
> ...and instead doing something like this:
>
> extern void pgstat_report_waiting(char *reason);
>
> ...an
On Thu, Nov 25, 2010 at 9:00 AM, Alvaro Herrera
wrote:
> Excerpts from Robert Haas's message of mar nov 23 00:08:54 -0300 2010:
>
>> How about publishing additional details to pg_stat_activity via
>> pgstat_report_waiting()?
>
> I'm not sure what you mean here. Are you suggesting we should create
On 25.11.2010 16:16, Itagaki Takahiro wrote:
On Thu, Nov 25, 2010 at 22:03, Heikki Linnakangas
wrote:
I propose the attached API instead. This has a clear separation between plan
and execution.
The APIs seem to be cleaner. The previous ones might be too straight
implementation of the SQL sta
Excerpts from Tom Lane's message of lun nov 22 20:51:09 -0300 2010:
> Alvaro Herrera writes:
> > A much more common ocurrence is tuple locks. We block in an Xid in that
> > case; and this has been a frequent question in the mailing lists and
> > IRC.
>
> > I think it would be very nice to be abl
On Thu, Nov 25, 2010 at 22:03, Heikki Linnakangas
wrote:
> I propose the attached API instead. This has a clear separation between plan
> and execution.
The APIs seem to be cleaner. The previous ones might be too straight
implementation of the SQL standard.
But I have some questions about the ne
Excerpts from Robert Haas's message of mar nov 23 00:08:54 -0300 2010:
> How about publishing additional details to pg_stat_activity via
> pgstat_report_waiting()?
I'm not sure what you mean here. Are you suggesting we should create a
new function with that name to report the reason for the lock
On Sat, Nov 20, 2010 at 2:04 AM, Tom Lane wrote:
> Alvaro Herrera writes:
>> Excerpts from Tom Lane's message of vie nov 19 12:25:13 -0300 2010:
>>> Yeah. You're adding a new fundamental state to the protocol; it's not
>>> enough to bury that in the description of a message format. I don't
>>>
On 25.11.2010 09:34, Shigeru HANADA wrote:
Attached is a patch that adds core functionality of SQL/MED. This
patch provides:
* new option HANDLER for FOREIGN DATA WRAPPER
* CREATE/ALTER DDLs are supported
* psql \dew command shows handler option too
* pg_dump can dump HANDLER option
*
Excerpts from Magnus Hagander's message of jue nov 25 05:46:49 -0300 2010:
> On Wed, Nov 24, 2010 at 23:54, Tom Lane wrote:
> > Its not so much the moderation load, as I don't like being blindsided by
> > commits that touch everything in sight. Finding out only when you try
> > to do git push (a
I checked against other parameter bindings and it looks like problem is
connected with oid=0.
In those cases:
1. Executing statement with parameter sent as varchar, int, long, with
text and binary format is ok.
2. Executing statement with oid=0 fail always; I've sent parameter in text
mode (encode
On Tue, Nov 23, 2010 at 02:21, Tom Lane wrote:
> In any case I don't see an argument why warning on connection creation
> isn't sufficient.
I'll check all versions of dblink. truncate_identifier() will be called
with warn=false in all cases except dblink_coneect() -> createNewConnection().
--
>I've seen this as well be a performance issue, in particular with partitioned
>tables.
>Out of habit I now write functions that always cache the value of the function
>in
>a variable and use the variable in the actual query to avoid this particular
>"gotcha".
subquery may be used to cache cons
Haven't really gotten much further, but an interesting note: the named
/ unnamed prepared statement and portal stuff seems to be a red
herring. I can add a name to the portal, or move to an unnamed
prepared statement, and I still see the same thing. Which is
interesting, since that's not what Rados
Hi,
2010/11/25 Markus Wanner :
> Eliot,
>
> On 11/23/2010 09:43 PM, Eliot Gable wrote:
>> I know there has been a lot of talk about replication getting built into
>> Postgres and I know of many projects that aim to fill the role. However,
>> I have not seen much in the way of a serious attempt at
Eliot,
On 11/23/2010 09:43 PM, Eliot Gable wrote:
> I know there has been a lot of talk about replication getting built into
> Postgres and I know of many projects that aim to fill the role. However,
> I have not seen much in the way of a serious attempt at multi-master
> write scaling.
Postgres-
On Thu, 25 Nov 2010 17:12:44 +0900
Shigeru HANADA wrote:
> Attached is a patch that adds file_fdw, FDW which reads records from
> files on the server side, as a contrib module. This patch is based on
> "SQL/MED core functionality" patch.
>
> [SQL/MED - core functionality]
> http://archives.postg
On Wed, Nov 24, 2010 at 23:54, Tom Lane wrote:
> Magnus Hagander writes:
>> On Wed, Nov 24, 2010 at 23:45, Tom Lane wrote:
>>> Alvaro Herrera writes:
This was stuck in the moderation queue because of message size limit (30
kB). Is it worth increasing that value?
>>>
>>> Evidently we
Hi, hackers,
Attached is a patch that adds postgresql_fdw, FDW which retrieves
tuples from external PostgreSQL server, as a contrib module. This
patch is based on "SQL/MED core functionality" patch.
[SQL/MED - core functionality]
http://archives.postgresql.org/pgsql-hackers/2010-11/msg01698.php
Hi, hackers,
Attached is a patch that adds file_fdw, FDW which reads records from
files on the server side, as a contrib module. This patch is based on
"SQL/MED core functionality" patch.
[SQL/MED - core functionality]
http://archives.postgresql.org/pgsql-hackers/2010-11/msg01698.php
File_fdw c
On Wed, Nov 24, 2010 at 1:27 PM, Fujii Masao wrote:
> Hi,
>
> http://archives.postgresql.org/pgsql-hackers/2010-11/msg01303.php
>
> When I did unusual operations (e.g., suspend bgwriter by gdb,
> pgbench -i and issue txid_current many times) on the master
> in order to try to reproduce the above H
69 matches
Mail list logo
