Re: Savepoint and prepared transactions - Mailing list pgsql-general

From Scott Marlowe
Subject Re: Savepoint and prepared transactions
Date
Msg-id g2xdcc563d11005050547wc5d393c9t796a7a4e0215d728@mail.gmail.com
Whole thread Raw
In response to Re: Savepoint and prepared transactions  (Konstantin Izmailov <pgfizm@gmail.com>)
Responses Re: Savepoint and prepared transactions  (Konstantin Izmailov <pgfizm@gmail.com>)
List pgsql-general
Should work.  I'm not sure 2.5 release savepoint is necessary.

On Wed, May 5, 2010 at 12:53 AM, Konstantin Izmailov <pgfizm@gmail.com> wrote:
> I expect that only changes on step 2.7 persisted in DB.
>
> On Tue, May 4, 2010 at 11:18 PM, Scott Marlowe <scott.marlowe@gmail.com>
> wrote:
>>
>> On Tue, May 4, 2010 at 11:52 PM, Konstantin Izmailov <pgfizm@gmail.com>
>> wrote:
>> > Hi,
>> > I'm working on a solution that utilizes 2 phase commit protocol (between
>> > SQL
>> > Server and PostgreSQL). Normally PostgreSQL statements sequense is:
>> > 1. START
>> > 2. <inserts, updates, etc.>
>> > 3. PREPARE TRANSACTION 'uuid'
>> > 4. COMMIT PREPARED 'uuid'
>> >
>> > What if on step 2 user application issues statements with SAVEPOINTs,
>> > e.g.
>> > 2.1. SAVEPOINT svp1
>> > 2.2. <inserts, updates, etc.>
>> > 2.3. SAVEPOINT svp2
>> > 2.4. <inserts, updates, etc.>
>> > 2.5. RELEASE SAVEPOINT svp2
>> > 2.6. ROLLBACK TO SAVEPOINT svp1
>> > 2.7. <inserts, updates, etc.>
>> >
>> > Is this allowed and safe to use?
>>
>> What are you expecting to happen?
>
>



--
When fascism comes to America, it will be intolerance sold as diversity.

pgsql-general by date:

Previous
From: Thom Brown
Date:
Subject: Re: alter table alter type CASCADE
Next
From: Scott Marlowe
Date:
Subject: Re: alter table alter type CASCADE