Re: Further pg_upgrade analysis for many tables

Stephen Frost <sfrost@snowman.net> writes:
> I'm curious if this is going to help with rollback's of transactions
> which created lots of tables..?  We've certainly seen that take much
> longer than we'd like, although I've generally attributed it to doing
> all of the unlink'ing and truncating of files.

If a single transaction creates lots of tables and then rolls back,
this patch won't change anything because we'll long since have
overflowed the eoxact list.  But you weren't seeing an O(N^2) penalty
in such cases anyway: that penalty came from doing O(N) work in each
of N transactions.  I'm sure you're right that you're mostly looking
at the filesystem cleanup work, which we can't do much about.
        regards, tom lane