Re: CLUSTER and MVCC

Heikki Linnakangas wrote:
> One potential issue I'm seeing is that if we rely on the unbroken chain
> starting from < OldestXmin, and that tuple isn't there because of a bug,
> for example, the later version of the tuple is skipped and the row is lost.

After further thought, I don't feel comfortable with the idea because of
the above loss of robustness.

I'm thinking of keeping an in-memory mapping of old and new tids of
updated tuples while clustering, instead. That means that cluster
requires a little bit of memory for each RECENTLY_DEAD updated tuple. In
the worst case that means that you run out of memory if there's too many
of those in the table, but I doubt that's going to be a problem in practice.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com