On Fri, 2009-05-29 at 14:54 -0400, Robert Haas wrote:
> On Fri, May 29, 2009 at 2:23 PM, Simon Riggs <simon@2ndquadrant.com> wrote:
> > Regrettably, the patch doesn't remove the problem it was supposed to
> > remove and I'm highlighting there is still risk of data loss. I suggest
> > that we don't change any docs, and carefully word or even avoid any
> > release note inclusion to avoid lulling people into stopping safety
> > measures.
>
> I think it's pretty clear that you and the OP are talking about two
> different problems. To quote Guillaume:
>
> "Yes, the problem is that before this change, even with a working
> replication and a clean shutdown, you still had to replicate the last
> WAL file by hand."
>
> I think that's a pretty legitimate complaint.
It's valid complaint, yes, but only for people that do this manually,
which is nobody I ever met, in *production*. (ymmv etc)
> You seem to think that this wasn't worth fixing...
And for them, it hasn't been completely fixed. That point was not made
by patch author or committer, leaving the impression it was now
completely safe, which, I truly regret to say, is not correct.
-- Simon Riggs www.2ndQuadrant.comPostgreSQL Training, Services and Support