On 12/23/14, 8:54 AM, Fabrízio de Royes Mello wrote:
> Right now a lot of people just work around this with things like DO blocks, but as mentioned elsewhere in the thread that fails for commands that can't be in a transaction. >
I use "dblink" to solve it. :-)
So... how about instead of solving this only for vacuum we create something generic? :) Possibly using Robert's background worker work?
Interesting idea.
But and what about the idea of improve the "--table" option from clients: vaccumdb and clusterdb?
Regards,
Fabrízio Mello
--
Fabrízio de Royes Mello Consultoria/Coaching PostgreSQL