On Thursday 29 May 2008 12:13:20 Bruce Momjian wrote:
> David Fetter wrote:
> > On Thu, May 29, 2008 at 11:58:31AM -0400, Bruce Momjian wrote:
> > > Josh Berkus wrote:
> > > > Publishing the XIDs back to the master is one possibility. We
> > > > also looked at using "spillover segments" for vacuumed rows, but
> > > > that seemed even less viable.
> > > >
> > > > I'm also thinking, for *async replication*, that we could simply
> > > > halt replication on the slave whenever a transaction passes minxid
> > > > on the master. However, the main focus will be on synchrounous
> > > > hot standby.
> > >
> > > Another idea I discussed with Tom is having the slave _delay_
> > > applying WAL files until all slave snapshots are ready.
> >
> > Either one of these would be great, but something that involves
> > machines that stay useless most of the time is just not going to work.
>
> Right, the ultimate target is to have the slave be read-only, but we
> need to get the streaming of WAL logs done first.
>
I would have thought the read only piece would have been more important than
the synchronous piece. In my experience readable slaves is the big selling
point in both Oracle and MySQL's implementations, and people are not nearly
as concerned if there is a small asynchronous window.
It would also be easier to implement on some level; we have already solved the
asynchronus wal shipping problem, so we would just need to solve the
read-only bits. For synchronus hot standby, you have to solve both the
synchronus shipping and the read-only bits. Seems like more work with less
upside that read-only slaves vs. pitr warm standby we have now.
Interesting that core views this differently.
--
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL