Horton

Re: pg_upgrade and statistics

Euler Taveira de Oliveira <euler@timbira.com>

View thread
From: Euler Taveira <euler@timbira.com>
To: Bruce Momjian <bruce@momjian.us>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Kevin Grittner <Kevin.Grittner@wicourts.gov>, Peter Eisentraut <peter_e@gmx.net>, Daniel Farina <daniel@heroku.com>, Greg Stark <stark@mit.edu>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2012-03-14T04:03:26Z
Lists: pgsql-hackers 
On 13-03-2012 21:34, Bruce Momjian wrote:
> It might be a solution for cases where we don't modify it.  I frankly am
> worried that if we copy over statistics even in ASCII that don't match
> what the server expects, it might lead to a crash, which has me back to
> wanting to speed up vacuumdb.
> 
That was discussed in another thread some time ago [1]. Adopting a hack
solution is not the way to go. It could lead to bad consequences in a near future.

For 9.2, we could advise users to divide the ANALYZE step into
ANALYZE-per-table steps and run them all in parallel. This ANALYZE-per-table
ranking could be accomplished using a simple approach like '... row_number()
OVER (ORDER BY pg_relation_size(oid) ... WHERE row_number % n = x' (tip stolen
from Simon's book).


[1] http://archives.postgresql.org/message-id/4F10A728.7090403@agliodbs.com


-- 
   Euler Taveira de Oliveira - Timbira       http://www.timbira.com.br/
   PostgreSQL: Consultoria, Desenvolvimento, Suporte 24x7 e Treinamento