Re: Repeatable read and serializable transactions see data committed after tx start

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: Álvaro Hernández Tortosa <aht(at)8kdata(dot)com>
Cc: "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Repeatable read and serializable transactions see data committed after tx start
Date: 2014-11-07 20:48:36
Message-ID: CA+TgmoZAj2e0irM7DtrtKam751BVvLmJqi8=pSO=59R3-5WwEQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Nov 5, 2014 at 7:17 PM, Álvaro Hernández Tortosa <aht(at)8kdata(dot)com> wrote:
>>> Given a transaction started with "BEGIN.... (REPEATABLE READ |
>>> SERIALIZABLE)", if a concurrent session commits some data before *any*
>>> query
>>> within the first transaction, that committed data is seen by the
>>> transaction. This is not what I'd expect.
>>
>> I think the problem is with your expectation, not the behavior.
>
> But my expectation is derived from the documentation:
>
> "The Repeatable Read isolation level only sees data committed before the
> transaction began;"

Yes, that's inaccurate. We should fix it.

(But we should not change the behavior, because I'm pretty sure that
the villagers would show up with pitchforks when various
currently-harmless scenarios caused massive database bloat.)

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

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2014-11-07 20:52:04 Re: B-Tree index builds, CLUSTER, and sortsupport
Previous Message Robert Haas 2014-11-07 20:40:37 Re: On partitioning