Re: Implementing incremental backup - Mailing list pgsql-hackers

From Andres Freund
Subject Re: Implementing incremental backup
Date
Msg-id 20130622140851.GA1254@alap2.anarazel.de
Whole thread Raw
In response to Re: Implementing incremental backup  (Cédric Villemain <cedric@2ndquadrant.com>)
List pgsql-hackers
On 2013-06-22 15:58:35 +0200, Cédric Villemain wrote:
> > A differential backup resulting from a bunch of WAL between W1 and Wn
> > would help to recover much faster to the time of Wn than replaying all
> > the WALs between W1 and Wn and saves a lot of space.
> > 
> > I was hoping to find some time to dig around this idea, but as the
> > subject rose here, then here are my 2¢!
> 
> something like that maybe :
> ./pg_xlogdump -b \
>     ../data/pg_xlog/000000010000000000000001 \   
>     ../data/pg_xlog/000000010000000000000005| \
>     grep 'backup bkp' | awk '{print ($5,$9)}'

Note that it's a bit more complex than that for a number of reasons:
* we don't log full page images for e.g. new heap pages, we just set the XLOG_HEAP_INIT_PAGE flag on the record
* there also are XLOG_FPI records
* How do you get a base backup as the basis to apply those to? You need it to be recovered exactly to a certain
point...

But yes, I think something can be done in the end. I think Heikki's
pg_rewind already has quite a bit of the required logic.

Greetings,

Andres Freund

-- Andres Freund                       http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services



pgsql-hackers by date:

Previous
From: Cédric Villemain
Date:
Subject: Re: Implementing incremental backup
Next
From: Andres Freund
Date:
Subject: Re: Possible bug in CASE evaluation