Jayadevan M wrote:
> I went through the documentation on replication. and am trying to
understand the various options
> available.
>
> Does replication available in PostgreSQL always use WAL (i.e
irrespective of whether it is
> synchronous/asynchronous, whether it is using archived logs or 'live'
logs or streaming)?
Yes as far as replication solutions included in core are concerned.
There are other replication solutions (like Slony) using triggers.
> I understand that we can use WALs in 3 modes
>
> 1) Use the archived WALs - results in the slave being
significantly behind the master. This is
> set up using the archive_command
>
> 2) Use entire WAL segments
I am not sure what the difference between 1) and 2) is.
There will be a significant lag, yes.
> 3) Use 'streaming' , i.e. use WAL files before they are complete
- in this case the slave is
> almost uptodate
Right.
> If I set up a slave using the archived WAls and set hot_standby on in
the slave, will I be able to
> read from the slave?
Yes.
Yours,
Laurenz Albe
