Re: ALTER TABLE lock strength reduction patch is unsafe

On March 4, 2014 8:39:55 PM CET, Simon Riggs <simon(at)2ndQuadrant(dot)com> wrote:
>On 4 March 2014 16:27, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> writes:
>>> One concern is schema changes that make a dump unrestorable, for
>>> instance if there's a foreign key relationship between tables A and
>B,
>>
>> Yeah. Ideally, what pg_dump would produce would be a consistent
>snapshot
>> of the database state as of its transaction snapshot time. We have
>always
>> had that guarantee so far as user data was concerned, but it's been
>shaky
>> (and getting worse) so far as the database schema is concerned. What
>> bothers me about the current patch is that it's going to make it a
>whole
>> lot more worse.
>
>While thinking this through it occurs to me that there is no problem at
>all.
>
>Your earlier claim that the dump is inconsistent just isn't accurate.
>We now have MVCC catalogs, so any dump is going to see a perfectly
>consistent set of data plus DDL. OK the catalogs may change AFTER the
>snapshot was taken for the dump, but then so can the data change -
>that's just MVCC.
>
>I was going to add an option to increase lock level, but I can't see
>why you'd want it even. The dumps are consistent...

Mvcc scans only guarantee that individual scans are consistent, not that separate scans are. Each individual scan takes a new snapshot if there's been ddl.

Andres

--
Please excuse brevity and formatting - I am writing this on my mobile phone.

Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services