Thom Brown wrote:
> On 14 June 2011 19:42, Bruce Momjian <bruce@momjian.us> wrote:
> > In pg_upgrade, check there are no prepared transactions.
>
> Out of curiosity, why would you check the new cluster for prepared transactions?
Yes, I was debating mentioning why I did that. First, the new cluster
should be empty, so there is a procedural issue that if it isn't, they
probably did something wrong. Second, I am worried that replying a
prepared transaction, now that transaction ids are now different
(remember we froze _everthing_) might not work, and I don't want to try
to find out that it works now and doesn't work later, to safety says to
just disallow it.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +