Re: Fully-automatic streaming replication failover when master dies?

On Mon, Jan 27, 2014 at 12:35 PM, Sameer Kumar <sameer(dot)kumar(at)ashnik(dot)com>wrote:

>
> On Mon, Jan 27, 2014 at 11:24 AM, Michael Paquier <
> michael(dot)paquier(at)gmail(dot)com> wrote:
>
>> PostgreSQL supports synchronous multi-master, MongoDB supports write
>>>> concern, but this causes a performance penalty).
>>>
>>> Anyways I doubt that "PostgreSQL supports synchronous multi-master"
>>>
>> Postgres core taken as such does not support multi-master. The fork of
>> PostgreSQL called Postgres-XC somewhat does, for OLTP applications.
>>
>
> I know about PostgresXC, but I thought it is distributed database (similar
> to shards of mongoDB). [Though if I am correct there could be tables which
> are shared across different nodes, but that is not the best way of
> utilizing features of PostgresXC] I think it is not apt to call it
> "synchronous" (since there is no replication happening).
>
XC supports table sharding with distribution of data using a column key, or
even data replication. Table definitions and schemas are the same across
all the nodes, it is just that the data is present in a portion of the
nodes (Datanodes more precisely) of the cluster. Operations are done
*synchronously* through the cluster using what is called a GTM (global
transaction manager) that is unique and feeds all the other nodes with
globally-consistent transaction IDs and snapshots. GTM SPOF is solved with
the presence of a Standby.
--
Michael