Thread: drop-index-concurrently-1 on master fails at serializable

drop-index-concurrently-1 on master fails at serializable

From
Kevin Grittner
Date:
*** /home/kgrittn/pg/master/src/test/isolation/expected/drop-index-concurrently-1.out   2013-07-10 14:58:41.641205557
-0500
--- /home/kgrittn/pg/master/src/test/isolation/results/drop-index-concurrently-1.out    2013-10-07 12:20:49.269277851
-0500
***************
*** 30,40 ****
  id             data          
 
  34             34            
- 134            34            
  step selects: EXECUTE getrow_seq;
  id             data          
 
  34             34            
- 134            34            
  step end: COMMIT;
  step drop: <... completed>
--- 30,38 ----

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



Re: drop-index-concurrently-1 on master fails at serializable

From
Kevin Grittner
Date:
Kevin Grittner <kgrittn@ymail.com> wrote:

> [ isolation test failed at snapshot-based isolation levels ]

Fix pushed, that looks for the right results based on isolation level.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



Re: drop-index-concurrently-1 on master fails at serializable

From
Andres Freund
Date:
On 2013-10-08 15:01:26 -0700, Kevin Grittner wrote:
> Kevin Grittner <kgrittn@ymail.com> wrote:
> 
> > [ isolation test failed at snapshot-based isolation levels ]
> 
> Fix pushed, that looks for the right results based on isolation level.

Hm, given what we're trying to test here, wouldn't it be better to
explicitly use READ COMMITTED?

Greetings,

Andres Freund

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



Re: drop-index-concurrently-1 on master fails at serializable

From
Kevin Grittner
Date:
Andres Freund <andres@2ndquadrant.com> wrote:
> On 2013-10-08 15:01:26 -0700, Kevin Grittner wrote:
>> Kevin Grittner <kgrittn@ymail.com> wrote:
>>
>> > [ isolation test failed at snapshot-based isolation levels ]
>>
>> Fix pushed, that looks for the right results based on isolation level.
>
> Hm, given what we're trying to test here, wouldn't it be better to
> explicitly use READ COMMITTED?

I thought about that approach, but it seemed better to make sure
that things didn't get broken at any isolation level by patches
dealing with DROP INDEX CONCURRENTLY.  If you're sure that could
never happen, we could save a few dozen lines of isolation test
code.

It's not like READ COMMITTED will never get tested -- I would bet
that upwards of 99% of the make installcheck-world runs or make
installcheck -C src/test/isolation runs are at that isolation
level.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company