Re: Concurrent ALTER SEQUENCE RESTART Regression

On Wed, Apr 26, 2017 at 9:51 PM, Michael Paquier
<michael(dot)paquier(at)gmail(dot)com> wrote:
> On Thu, Apr 27, 2017 at 10:12 AM, Andres Freund <andres(at)anarazel(dot)de> wrote:
>> On 2017-04-26 21:07:20 -0400, Peter Eisentraut wrote:
>>> One thing we could do, since all catalog updates now go through
>>> CatalogTupleUpdate(), is not use simple_heap_update() there but do the
>>> heap_update() directly and provide a better user-facing error message.
>>
>> I think it's unacceptable to regress with an error message here. I've
>> seen sequence DDL being used while concurrent DML was onging in a number
>> of production use cases, and just starting to error out instead of
>> properly blocking doesn't seem acceptable to me.
>
> +1'ing on this argument.

+1 from me, too. The fact that we have some other places where we get
"ERROR: tuple concurrently updated" is an awfulness that nobody's
gotten around to fixing, not an excuse to regress cases that were
previously working properly (never mind the other breakage reported
downthread). If this can't be fixed the whole patch should be ripped
out, IMHO.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company