While researching a problem reported on the -general list by a user who
lost a disk containing his index tablespace, I ran into something, but I'm
not sure is a serious bug or just an inconsistency in how \d shows tables.
Here are the steps I took.
1. Create a new database 'MYDB' and connect to
I spent a significant chunk of my time last week, and also a whole lot
of machine time, trying to evaluate the effectiveness of flushing CLOG
pages to disk in the background. Simon made the last effort in this
area:
http://archives.postgresql.org/pgsql-hackers/2012-01/msg00571.php
...but we were
On Fri, May 4, 2012 at 12:41 PM, Armando
wrote:
> Hi everybody.
>
> First of all I have to thank you for your wonderful job! PostgreSQL rocks!
>
> I am writing you because I am interested in understanding some specifics
> related
> to PostgreSQL internals. More precisely, I am investigating the r
On 7 May 2012 20:06, Tom Lane wrote:
> Simon Riggs writes:
>> It also leaves the situation that we have a catalog view called
>> pg_stat_bgwriter that would be accessing "checkpointer" things. That's
>> really the thorny one that I wasn't sure how to handle. Good example
>> of why we shouldn't ex
Simon Riggs writes:
> It also leaves the situation that we have a catalog view called
> pg_stat_bgwriter that would be accessing "checkpointer" things. That's
> really the thorny one that I wasn't sure how to handle. Good example
> of why we shouldn't expose internals too much.
Yeah, that's a bit
On 7 May 2012 19:44, Tom Lane wrote:
> Simon Riggs writes:
>> On 7 May 2012 18:09, Tom Lane wrote:
>>> I also notice that the separate-checkpointer patch failed to rename
>>> assorted things like BgWriterCommLock, BgWriterRequest,
>>> BgWriterShmemStruct, which are all 100% inappropriately named
Simon Riggs writes:
> On 7 May 2012 18:09, Tom Lane wrote:
>> I also notice that the separate-checkpointer patch failed to rename
>> assorted things like BgWriterCommLock, BgWriterRequest,
>> BgWriterShmemStruct, which are all 100% inappropriately named now.
>> And it still contains various obsol
On 7 May 2012 18:09, Tom Lane wrote:
> Peter Geoghegan writes:
>> This latest revision also covers the checkpointer. The code for that
>> is far simpler than that for the WAL Writer, so it doesn't
>> particularly feel like I'm pushing my luck by slipping that into
>> something to be slipped in.
>
Magnus Hagander writes:
> On Mon, May 7, 2012 at 7:18 PM, Robert Haas wrote:
>> Since the following hunk is repeated 3x, maybe it should be stuffed
>> into a function that is then called in three places:
> I considered it trivial enough not to do that for it. I can perhaps be
> convinced otherwi
On Sat, May 5, 2012 at 9:03 AM, Bruce Momjian wrote:
> On Fri, May 04, 2012 at 08:46:28PM +0300, Peter Eisentraut wrote:
>> On tor, 2012-05-03 at 15:47 -0400, Bruce Momjian wrote:
>> > Peter, where are we on this?
>>
>> I hadn't received any clear feedback, but if no one objects, I can
>> commit i
On Mon, May 7, 2012 at 7:18 PM, Robert Haas wrote:
> On Mon, May 7, 2012 at 12:39 PM, Magnus Hagander wrote:
>> Makes sense, will change and commit.
>
> Since the following hunk is repeated 3x, maybe it should be stuffed
> into a function that is then called in three places:
I considered it triv
On Mon, May 7, 2012 at 12:39 PM, Magnus Hagander wrote:
> Makes sense, will change and commit.
Since the following hunk is repeated 3x, maybe it should be stuffed
into a function that is then called in three places:
+ if (IsTransactionState())
+ ereport(COMMER
Peter Geoghegan writes:
> This latest revision also covers the checkpointer. The code for that
> is far simpler than that for the WAL Writer, so it doesn't
> particularly feel like I'm pushing my luck by slipping that into
> something to be slipped in.
Well ... maybe, or maybe not, or maybe you a
On Mon, May 7, 2012 at 5:15 PM, Tom Lane wrote:
> Magnus Hagander writes:
>> Any further suggestoins for which codes to use? If not, I think I'm
>> going to commit the patch as I had it, because it's not any worse than
>> what we had before (but fixes the annoying messages), and we can
>> always
On Sat, May 5, 2012 at 12:41 PM, Fujii Masao wrote:
> On Sat, Apr 28, 2012 at 4:00 AM, Robert Haas wrote:
>> On Fri, Apr 27, 2012 at 2:48 PM, Tom Lane wrote:
>>> I'm not necessarily opposed to commandeering the name "smart" for the
>>> new behavior, so that what we have to find a name for is the
Magnus Hagander writes:
> Any further suggestoins for which codes to use? If not, I think I'm
> going to commit the patch as I had it, because it's not any worse than
> what we had before (but fixes the annoying messages), and we can
> always revisit the actual errorcodes later.
I'm still a bit u
On Sun, May 6, 2012 at 9:04 AM, Peter Eisentraut wrote:
> I think a few more things could removed/simplified after the recent
> round of port removal:
>
> - Remove definition of offsetof() in c.h
I see no particular virtue to getting rid of this.
> - (Side point, the definition of endof() in the
FYI, I am planning to complete the 9.2 beta release notes by this
Wednesday night, America-time, so developers will have Thursday to make
adjustments before we ship the release notes as part of the beta.
I wanted to complete them sooner, but I also wanted to be current on
email before I started.
On Tue, Apr 24, 2012 at 10:55 PM, Noah Misch wrote:
> A key barrier to migrations from trigger-based replication to WAL-based
> replication is the lack of temporary tables under hot standby. I'd like to
> close that gap; the changes needed will also reduce the master-side cost of
> temporary tabl
On Thu, May 3, 2012 at 9:26 PM, Magnus Hagander wrote:
> On Thu, May 3, 2012 at 7:48 PM, Tom Lane wrote:
>> Magnus Hagander writes:
>>> Heh - we already used ERRCODE_CONNECTION_FAILURE on the errors in
>>> copy.c. Since COPY can only happen when there is a transaction
>>> (right?), I just change
Fujii Masao wrote:
>>> I'm not necessarily opposed to commandeering the name "smart" for the
>>> new behavior, so that what we have to find a name for is the old "smart"
>>> behavior. How about
>>>
>>> slow - allow existing sessions to finish (old "smart")
>>> smart - allow exis
21 matches
Mail list logo