Re: About the structure of WAL Files. - Mailing list pgsql-hackers

From Alvaro Herrera
Subject Re: About the structure of WAL Files.
Date
Msg-id 20060315135230.GI5222@surnet.cl
Whole thread Raw
In response to Re: [PATCHES] About the structure of WAL Files.  (Simon Riggs <simon@2ndquadrant.com>)
List pgsql-hackers
Simon Riggs wrote:
> On Wed, 2006-03-15 at 09:12 -0400, Alvaro Herrera wrote:
> > Charlie Wang wrote:
> > > I want to write a function to expose the content of WAL Files as streams,
> > > and then send to somewhere else, record by record. 
> > 
> > Re: the WAL records, most likely they are useless outside the server
> > that generated them, because they refer to relations using Oids, and to
> > specific page addresses which are probably not going to be exactly the
> > same anywhere else (consider VACUUM commands running at different
> > times).
> 
> Yet it should be pointed out that there is a commercial product that
> says it uses this information to provide a replication system for
> PostgreSQL, so presumably that problem can be solved...

Are you referring to Mammoth Replicator?  Let me point out that it
doesn't really use the WAL.

Now that I think of it, maybe the problem _can_ be solved.  Consider
e.g. the WAL-dispatcher resolving the relation Oids to relnames before
shipping, and the WAL-receiver resolving that back to a local Oid before
"applying".  And it only needs to get the tuple contents; it can ignore
the position, and it can resolve the indexing by itself.

-- 
Alvaro Herrera                                http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support


pgsql-hackers by date:

Previous
From: Simon Riggs
Date:
Subject: Re: [PATCHES] About the structure of WAL Files.
Next
From: "Jonah H. Harris"
Date:
Subject: Re: [PATCHES] About the structure of WAL Files.