From: | Marty Scholes <marty(at)outputservices(dot)com> |
---|---|
To: | pgsql-hackers(at)postgresql(dot)org |
Cc: | Marty Scholes <marty(at)outputservices(dot)com>, josh(at)agliodbs(dot)com |
Subject: | Re: tablespaces and DB administration |
Date: | 2004-05-28 15:52:13 |
Message-ID: | 40B7602D.9090708@outputservices.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Josh Berkus wrote:
> The whole point of tablespaces is to
> allow placing individual tables and indexes
> on seperate volumes.
That was one reason. I seem to recall several more:
* Putting data on cost appropriate media
Mentioned previously in this thread
* Balancing I/O across spindles
Also mentioned previously, many times
* Inode stability
The free space bitmap never changes during production. The inode table
never changes during production (save for last modified and last
accessed timestamps). This makes the filesystem VERY likely to be
functional in case of a crash. Jounaled file systems have reduced the
need for this.
* Device independence
Since a tablespace needs only some "file names" and the ability to
lseek() to any point in a file, a tablespace file can be a file on the
UFS (or its variantes), a disk partition, or even a tape drive device.
From | Date | Subject | |
---|---|---|---|
Next Message | Philip Warner | 2004-05-28 15:58:46 | Re: pg_dump --comment? |
Previous Message | Alvaro Herrera | 2004-05-28 15:46:07 | Re: Nested xacts: looking for testers and review |