On 04/29/2014 02:18 AM, 德哥 wrote:
> HI,
> This is so bad in pg_dump use, when a database so big.
> because pg_dump is also use repeatable read isolation. and when pg_dump backup database , the database will bloat .
Yep. One approach is to take a filesystem-level backup (ie. with
pg_start/stop_backup() or pg_basebackup), start a second server from
that backup, and run pg_dump against that.
> Can we optimize it?
Not easily. I'd love to do something about it, but it's going to be a
long project.
- Heikki