Re: Reconstructing transaction content after the fact

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
Cc: Lionel Bouton <lionel(dot)bouton(at)jtek(dot)fr>, pgsql-general(at)lists(dot)postgresql(dot)org
Subject: Re: Reconstructing transaction content after the fact
Date: 2020-09-02 19:35:41
Message-ID: 427489.1599075341@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> writes:
> On 2020-Sep-02, Tom Lane wrote:
>> See VACUUM FREEZE. You couldn't hide connections immediately after
>> insertion, but if the idea is to sanitize every so often, it'd help.

> Starting with 9.4 which introduced the use of HEAP_XMIN_FROZEN
> combination in infomask to replace rewriting the xmin value proper, the
> Xids will be preserved by freezing, so that won't help.

Ah, right.

> One option to hide the xids might be to recreate the tables every once
> in a while, with something like

Hmm. Expensive, but if you were willing to make it even more expensive,
you could also defeat the tuple-ordering attacks mentioned upthread:

CREATE TABLE votes_copy AS SELECT * FROM votes ORDER BY random();

regards, tom lane

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Peter Eisentraut 2020-09-03 05:01:17 Re: SSL between Primary and Seconday PostgreSQL DBs
Previous Message Alvaro Herrera 2020-09-02 19:26:20 Re: Reconstructing transaction content after the fact