On 19/10/10 13:16, Josh Berkus wrote:
> Robert asked me to write this up, so here it is.
>
> It is critical that we make replication easier to set up, administrate
> and monitor than it currently is. In my conversations with people,
> this is more important to our users and the adoption of PostgreSQL
> than synchronous replication is.
>
> First, I'm finding myself constantly needing to tutor people on how to
> set up replication. The mere fact that it requires a minimum 1-hour
> class to explain how to use it, or a 10-page tutoral, tells us it's
> too complex. As further evidence, Bruce and I explained binary
> replication to several MySQL geeks at OpenSQLCamp last weekend, and
> they were horrified at the number and complexity of the steps
> required. As it currently is, binary replication is not going to win
> us a lot of new users from the web development or virtualization world.
>
+1
I've been having the same experience - how to set this up and do
failover and failback etc occupies quite a bit of time in courses I've
been teaching here in NZ and Australia. Having this whole replication
business much simpler is definitely the way to go.
A good example of how simple it can be is mongodb, where it is
essentially one command to setup a 2 replica system with a voting arbiter:
$ mongo> rs.initiate( { _id : "replication_set0", members : [ { _id : 0, host :
"192.163,2,100"}, { _id : 1, host : "192.168.2.101" }, { _id : 2, host :
"192.168.2.103",arbiterOnly : true } ] }
)