On 5 Srpen 2011, 9:00, Vivekkumar Pandey wrote:
> Hi all,
>
> I have a cluster database with a master and slave ,
> the size of master database is very high reltive to slave
> while both have approximatly same data.
What version of PostgreSQL is this, what kind of cluster, and what do you
mean by "approximately"?
With xlog-based replication (file shipping or streaming replication), the
relations should have exactly the same size, but they should also contain
exactly the same data.
That suggests you're using something else to build the cluster (e.g. slony
or something like that). In that case the size difference may be simply
due to data differences or dead tuples. VACUUM FULL should compact the
dead tuples, but it's not a cheap command (takes exclusive locks, time and
memory).
Tomas