Re: UNDO and in-place update

From: Thomas Kellerer <spam_eater(at)gmx(dot)net>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: UNDO and in-place update
Date: 2016-11-24 17:09:04
Message-ID: 1480007344350-5931844.post@n3.nabble.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

> FWIW, while this is basically true, the idea of repurposing UNDO to be
> usable for MVCC is definitely an Oracleism. Mohan's ARIES paper says
> nothing about MVCC.
> For snapshot isolation Oracle has yet a *third* copy of the data in a
> space called the "rollback segment(s)".

UNDO and rollback segment are the same thing in Oracle.

In older versions it was just called "rollback segment" I think it started
with Oracle 10g that they called it UNDO.

> Fwiw, Oracle does not use the undo log for snapshot fetches.
> It's used only for transaction rollback and recovery.

As UNDO and "rollback" are the same they do use the UNDO information for
MVCC:

http://docs.oracle.com/database/121/CNCPT/consist.htm#GUID-00A3688F-1219-423C-A5ED-4B8F25BEEAFB__BABFDBAJ

--
View this message in context: http://postgresql.nabble.com/UNDO-and-in-place-update-tp5931575p5931844.html
Sent from the PostgreSQL - hackers mailing list archive at Nabble.com.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2016-11-24 17:23:28 Re: UNDO and in-place update
Previous Message Tom Lane 2016-11-24 16:51:24 Re: [bugfix] commit timestamps ERROR on lookup of FrozenTransactionId