Re: Sync Rep Design

On Sun, 2011-01-02 at 12:13 -0800, MARK CALLAGHAN wrote:
> On Thu, Dec 30, 2010 at 9:02 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> > <reads MySQL documentation>
> >
> > I see now that you've tried to design this feature in a way that is
> > similar to MySQL's offering, which does have some value. But it
> > appears to me that the documentation you've written here is
> > substantially similar to the MySQL 5.5 reference documentation. That
> > could get us into a world of legal trouble - that documentation is not
> > even open source, let alone BSD.
> >
> > http://dev.mysql.com/doc/refman/5.5/en/replication-semisync.html

Sorry, only just read that bit. Are they that similar? My docs are about
3 times longer and cover all sorts of things. I didn't intentionally
copy anything, but that doesn't really matter, what matters is that if
you think they are similar, legal people might. I've only read the URL
above, not the other links from it.

Robert, Can you identify which paragraphs need to be re-written? I won't
argue, I will just rewrite them or delete them and start afresh. Thanks
for being eagle-eyed.

> The docs originate from work done by my former team at Google. The
> content license on this is CC 3.0 BY-SA, so I don't think that should
> be a concern.
> http://code.google.com/p/google-mysql-tools/wiki/SemiSyncReplication
> http://code.google.com/p/google-mysql-tools/wiki/SemiSyncReplicationDesign

I guess that gets us off the hook a little bit, but not far enough for
my liking. Thanks for trying to save me!

> >From http://dev.mysql.com/doc/refman/5.5/en/replication-semisync.html)
> the MySQL docs don't mention that other transactions can view the
> committed data on the master between steps 1 and 2. Is that possible
> in this case?

Other transactions on the master cannot read data until after the
confirmation it is on the sync standby.

> As described in the the MySQL docs, semi-sync has another benefit for
> some deployments. It rate limits busy clients to prevent them from
> creating replication lag between the primary and standby servers. I
> also provided the text for that
> (http://bugs.mysql.com/bug.php?id=57911) if you are concerned about
> copying.

Yeh, I'm aware of the effect, but I'm not really seeing slowing down the
master as a benefit, its more an implication of synchronicity.

--
Simon Riggs http://www.2ndQuadrant.com/books/
PostgreSQL Development, 24x7 Support, Training and Services