Alvaro Herrera wrote:
> So I'm going to apply your patch to both 8.4 and HEAD; we can always
> improve it later, I guess.
Thank you for your applying.
I think the fix is ugly, too. We need to introduce cleaner solution for 8.5.
Regards,
---
ITAGAKI Takahiro
NTT Open Source Software Center
-
Itagaki Takahiro wrote:
>
> Here is a patch to fix a bug in handling default values in reloptions.
> This fix should be applied to HEAD and 8.4.0.
>
> I used 'magic number -1' to propagate "not-specified" information to
> autovacuum process. It might look strange because the default value is
> ou
Itagaki Takahiro wrote:
>
> Alvaro Herrera wrote:
>
> > > We should remember whether a field was specified or not independntly.
> > I'm not sure what you are suggesting ...?
>
> Please imagine that:
>
> =# SET autovacuum_vacuum_scale_factor = 0.05;
> =# ALTER TABLE tbl SET (autovacuum_vacuum_t
Alvaro Herrera wrote:
> > We should remember whether a field was specified or not independntly.
> I'm not sure what you are suggesting ...?
Please imagine that:
=# SET autovacuum_vacuum_scale_factor = 0.05;
=# ALTER TABLE tbl SET (autovacuum_vacuum_threshold = 10);
AutoVacOpts.vacuum_threshol
Itagaki Takahiro wrote:
>
> Alvaro Herrera wrote:
>
> > we could use that to set a boolean
> > in StdRdOptions that indicated whether they are set in reloptions or are
> > default values.
>
> Do you mean we will have a boolean field *for each* field in StdRdOptions?
No, I was thinking in a sin
Here is a patch to fix a bug in handling default values in reloptions.
This fix should be applied to HEAD and 8.4.0.
I used 'magic number -1' to propagate "not-specified" information to
autovacuum process. It might look strange because the default value is
out of range of the reloption, but I thi
Alvaro Herrera wrote:
> we could use that to set a boolean
> in StdRdOptions that indicated whether they are set in reloptions or are
> default values.
Do you mean we will have a boolean field *for each* field in StdRdOptions?
We should remember whether a field was specified or not independntly
Itagaki Takahiro wrote:
> To fix the bug, each field in StdRdOptions should have "not-specified" flag.
> If not specified, autovacuum should use current GUC settings instead of
> reloptions. Is it possible to change the default values of reloptions
> to some magic number (-1 or so) ?
Ah. After c
Itagaki Takahiro wrote:
> PostgreSQL version: 8.4.0
> Operating system: all versions
>
> If we set FILLFACTOR to a table, autovacuum parameters in postgresql.conf
> will not affect to the table; default values are always used.
>
> In 8.4.0, we create StdRdOptions if a relation has some fields i
Itagaki Takahiro escreveu:
> In 8.4.0, we create StdRdOptions if a relation has some fields in
> pg_class.reloption. Then, unspecified fields are filled with the
> default values. The values are always hard-coded default values and
> don't reflect current GUC settings.
>
Hey, how I couldn't notice
PostgreSQL version: 8.4.0
Operating system: all versions
If we set FILLFACTOR to a table, autovacuum parameters in postgresql.conf
will not affect to the table; default values are always used.
In 8.4.0, we create StdRdOptions if a relation has some fields in
pg_class.reloption. Then, unspecifie
11 matches
Mail list logo
