Hi,
On Friday, December 16, 2011 01:50:03 PM Havasvölgyi Ottó wrote:
> A simple query "insert into mytable default values" on a single
> serial-column table also leaks. It can also produced by pgbench. On
> Windows and Linux too.
Youve mentioned that in several threads now. If you really think th
Hi,
A simple query "insert into mytable default values" on a single
serial-column table also leaks. It can also produced by pgbench. On
Windows and Linux too.
Best regards,
Otto
2011/12/12 Jochen Erwied
> Monday, December 12, 2011, 12:33:03 PM you wrote:
>
> > On 12.12.2011 13:16, Matteo Bec
Monday, December 12, 2011, 12:33:03 PM you wrote:
> On 12.12.2011 13:16, Matteo Beccati wrote:
>> Just to clarify, am I correct assuming that the issue does not affect
>> tables which have non-indexed inet fields?
> Hmm, I think it might also affect queries that do large merge joins on
> inet f
On 12.12.2011 13:16, Matteo Beccati wrote:
On 12/12/2011 09:29, Heikki Linnakangas wrote:
On 12.12.2011 08:26, Tom Lane wrote:
Andres Freund writes:
In 3b8161723c645853021b57330dd2ea0484ec6131 Heikki made DatumGetInetP
unpack
toasted values. Unfortunately the btree support functions for the
Hi Heikki,
On 12/12/2011 09:29, Heikki Linnakangas wrote:
> On 12.12.2011 08:26, Tom Lane wrote:
>> Andres Freund writes:
>>> In 3b8161723c645853021b57330dd2ea0484ec6131 Heikki made DatumGetInetP
>>> unpack
>>> toasted values. Unfortunately the btree support functions for the
>>> inet type
>>> di
On 12.12.2011 08:26, Tom Lane wrote:
Andres Freund writes:
In 3b8161723c645853021b57330dd2ea0484ec6131 Heikki made DatumGetInetP unpack
toasted values. Unfortunately the btree support functions for the inet type
didn't free memory which they have to do in contrast to about everything else.
I
Monday, December 12, 2011, 1:45:42 AM you wrote:
> Jochen, could you try the attached patch?
Patch is missing a variable define for 'r' in network_sup(). Fixed patch
attached.
Memory usage for insert ~1087MB - no change
Memory usage for index creation ~415MB - fixed.
Memory usage for select ~15
Andres Freund writes:
> In 3b8161723c645853021b57330dd2ea0484ec6131 Heikki made DatumGetInetP unpack
> toasted values. Unfortunately the btree support functions for the inet type
> didn't free memory which they have to do in contrast to about everything else.
> I fixed a few more functions than
Hi,
On Monday, December 12, 2011 12:45:23 AM Andres Freund wrote:
> On Sunday, December 11, 2011 10:15:29 PM Jochen Erwied wrote:
> > The following script can be used to show the behaviour:
> >
> > create database pgtest;
> > \c pgtest
> > create table test(var inet);
> > insert into test(var) '1
Hi,
On Sunday, December 11, 2011 10:15:29 PM Jochen Erwied wrote:
> The following script can be used to show the behaviour:
>
> create database pgtest;
> \c pgtest
> create table test(var inet);
> insert into test(var) '127.0.0.0'::inet+generate_series(0,256*256*256-1);
> create index test_ix on
After upgrading from 9.1.1 to 9.1.2, memory usage for some queries
increased to the amount that the OOM-killer on Linux terminated the server.
Further investigation show the problem to occur under multiple circumstances:
- copying data from an external file into a table with indexes
- creating in
11 matches
Mail list logo
