From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Francisco J Reyes <fran(at)natserv(dot)net> |
Cc: | Christoph Dalitz <christoph(dot)dalitz(at)hs-niederrhein(dot)de>, Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>, PG Mailing List <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Status of tablespaces |
Date: | 2003-01-28 16:20:33 |
Message-ID: | 13677.1043770833@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Francisco J Reyes <fran(at)natserv(dot)net> writes:
> I also wonder if the syntax of the current create DB won't have to be
> changed once tablespaces are created.
I'd be inclined to remove the WITH LOCATION option the instant we have
an adequate tablespace substitute. It's such an ugly, messy, unsafe
way to do things --- depending on environment variables is bad news.
There are things I'm willing to keep around in the name of backward
compatibility, but this isn't one of them.
It's probably true that some other WITH options would have to be added
to CREATE DATABASE (and/or ALTER DATABASE) instead. For instance, I'd
imagine that we'd want to be able to specify a default tablespace for
a database; perhaps separate defaults for its tables and indexes.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Francisco J Reyes | 2003-01-28 16:24:45 | Re: Status of tablespaces |
Previous Message | Ron Johnson | 2003-01-28 16:15:08 | Re: Using RSYNC for replication? |