Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Bruce Momjian
On Wed, Oct 19, 2016 at 01:25:33PM -0400, Tom Lane wrote: > Bruce Momjian writes: > > On Wed, Oct 19, 2016 at 09:44:05AM -0700, David G. Johnston wrote: > >> ​I think the theory of having all system tables except LO on SSD storage, > >> and > >> having LO on a less performant device, makes sense.

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Tom Lane
Bruce Momjian writes: > On Wed, Oct 19, 2016 at 09:44:05AM -0700, David G. Johnston wrote: >> ​I think the theory of having all system tables except LO on SSD storage, and >> having LO on a less performant device, makes sense. > OK, so is this a TODO item? According to https://www.postgresql.org

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Bruce Momjian
On Wed, Oct 19, 2016 at 09:44:05AM -0700, David G. Johnston wrote: > On Wed, Oct 19, 2016 at 9:29 AM, Bruce Momjian wrote: > > On Tue, Oct 18, 2016 at 04:51:54PM +0200, Andreas Joseph Krogh wrote: > >     > 2. Being able to move pg_largeobject to a different tablespace > >     >    *w

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Andreas Joseph Krogh
På onsdag 19. oktober 2016 kl. 18:44:24, skrev Tom Lane mailto:t...@sss.pgh.pa.us>>: Andreas Joseph Krogh writes: > På onsdag 19. oktober 2016 kl. 18:29:31, skrev Bruce Momjian   I think an open question is why you would not want to move the other >  system tables at the same time you move pg_l

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Andreas Joseph Krogh
På onsdag 19. oktober 2016 kl. 18:42:11, skrev Bruce Momjian mailto:br...@momjian.us>>: On Wed, Oct 19, 2016 at 06:33:55PM +0200, Andreas Joseph Krogh wrote: > På onsdag 19. oktober 2016 kl. 18:29:31, skrev Bruce Momjian >: > >     On Tue, Oct 18, 2016 at 04:51:54PM +0200, Andreas Joseph Krogh

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Tom Lane
Andreas Joseph Krogh writes: > På onsdag 19. oktober 2016 kl. 18:29:31, skrev Bruce Momjian > I think an open question is why you would not want to move the other > system tables at the same time you move pg_largeobject. > Are you saying that if I move all system-tables to the tablespace I mo

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread David G. Johnston
On Wed, Oct 19, 2016 at 9:29 AM, Bruce Momjian wrote: > On Tue, Oct 18, 2016 at 04:51:54PM +0200, Andreas Joseph Krogh wrote: > > > 2. Being able to move pg_largeobject to a different tablespace > > >*without* turning on system_table_mods. This is important for > > >people sto

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Bruce Momjian
On Wed, Oct 19, 2016 at 06:33:55PM +0200, Andreas Joseph Krogh wrote: > På onsdag 19. oktober 2016 kl. 18:29:31, skrev Bruce Momjian >: > > On Tue, Oct 18, 2016 at 04:51:54PM +0200, Andreas Joseph Krogh wrote: > >     > 2. Being able to move pg_largeobject to a different tablespace >

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Andreas Joseph Krogh
På onsdag 19. oktober 2016 kl. 18:29:31, skrev Bruce Momjian mailto:br...@momjian.us>>: On Tue, Oct 18, 2016 at 04:51:54PM +0200, Andreas Joseph Krogh wrote: >     > 2. Being able to move pg_largeobject to a different tablespace >     >    *without* turning on system_table_mods. This is important

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-19 Thread Bruce Momjian
On Tue, Oct 18, 2016 at 04:51:54PM +0200, Andreas Joseph Krogh wrote: > > 2. Being able to move pg_largeobject to a different tablespace > >    *without* turning on system_table_mods. This is important for > >    people storing LOTS of large-objects on separate > >    disks (non-SSD

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-18 Thread Andreas Joseph Krogh
På tirsdag 18. oktober 2016 kl. 16:26:37, skrev Euler Taveira < eu...@timbira.com.br >: On 18-10-2016 10:13, Andreas Joseph Krogh wrote: > From time to time pg_largeobject comes up as an issue with being > implemented as a system-catalog. >  Did you read the archive

Re: [HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-18 Thread Euler Taveira
On 18-10-2016 10:13, Andreas Joseph Krogh wrote: > From time to time pg_largeobject comes up as an issue with being > implemented as a system-catalog. > Did you read the archives [1]? > As I see it, there are 2 relevant use-cases for improving the situation: > 1. Being able to pg_dump *without*

[HACKERS] Move pg_largeobject to a different tablespace *without* turning on system_table_mods.

2016-10-18 Thread Andreas Joseph Krogh
Hi -hackers.   >From time to time pg_largeobject comes up as an issue with being implemented as a system-catalog.   As I see it, there are 2 relevant use-cases for improving the situation: 1. Being able to pg_dump *without* any LOs (think of it as    without the contents of pg_largeobject). This