http://www.slony.info/bugzilla/show_bug.cgi?id=134

--- Comment #9 from Christopher Browne <cbbrowne at ca.afilias.info> 2010-11-15 14:44:30 PST ---
(In reply to comment #8)
> I reviewed a3a2e4839f5cf21bfd076e30e916cd23112daf97 rebased to our current
> master (3cbbaba072b38527fa2ac7949ba680e7bd410a9a) with conflicts resolved.
> 
> A few comments
> 
> 1.  Upgrade script support.  As it stands today an upgrade from 2.0
> to 2.1 doesn't require re-installing the cluster.  This patch doesn't
> warrent (in my opinion) requiring that.  Something else in the future
> might but I would like to see an upgrade script add the truncate triggers
> to all replicated tables.

Good idea; I will rework it so that it tries to add the triggers, if possible.

> 2. I manually tested the functionality and it looked okay.  However run I tried
> the regression test you added I got the same 'result' against 8.3 and
> 9.1master.
> Since 8.3 doesn't support truncate triggers this confused me, it might just be
> my unfamiliarity with the regression tests.  I was expecting a it to say it
> failed with 8.3 (is that what we want?)

Hmm.  I'd expect the test to fail on 8.3, because tables would be empty there,
but not on the subscribers.

I don't think I tried running it on 8.3, because I expected it not to work
properly there.

> 3. This probably should also have a documentation update, limitations.sgml at a
> minimum to say we now support truncate in 8.4 and above.  We also might want to
> say somewhere that the truncate on the slave is a CASCADE.   This would be
> important to note for setups who have a different schema on their slave vs
> master. 

Yep, something needs to be put in there.  I think I'll either need to:
a) Merge docs from HEAD into this branch, or
b) Document as part of the merge into HEAD, as
c) Try and muddle docs in as-is won't work out very happily due to the lots of
upstream changes to the docs.

Probably b) is simplest.

> Other than that I'm happy with this for master.

-- 
Configure bugmail: http://www.slony.info/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.