Re: [HACKERS] Isolation of table creation

2015-12-11 Thread Andres Freund
On 2015-12-11 11:46:11 +0300, Alexander Korotkov wrote:
> I discovered interesting issue with PostgreSQL transaction isolation.
> When transaction is in repeatable read isolation level, I can't see table
> which was created after transaction obtained snapshot. But I can run DML
> statements with this table. See example below.

> Is it a bug?

No, that's pretty much expected. Parse analysis, planning use fresh
catalog snapshot, whereas psql's schema queries use the transaction
snapshot.

Andres


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers


[HACKERS] Isolation of table creation

2015-12-11 Thread Alexander Korotkov
Hackers,

I discovered interesting issue with PostgreSQL transaction isolation.
When transaction is in repeatable read isolation level, I can't see table
which was created after transaction obtained snapshot. But I can run DML
statements with this table. See example below.

Session 1
# begin transaction isolation level repeatable read;
BEGIN
# \dt
No relations found.
Session 2
# create table tmp (i int not null);
CREATE TABLE
# insert into tmp values (1);
INSERT 0 1
# \dt
No relations found.
# select * from tmp;
 i
---
(0 rows)

# insert into tmp values (2);
INSERT 0 1
# select * from tmp;
 i
---
 2
(1 row)
# commit;
COMMIT

Is it a bug?

--
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company