Re: How to REINDEX in high volume environments?

Justin Clift dijo: 

Hi,

> Ran a benchmark against the database, REINDEX'd the tables, VACUUM FULL
> ANALYZE'd, prepared to re-run the benchmark again and guess what?
> 
> The indexes were back on the original drive.

Yes, this is expected.  Same for CLUSTER.  They create a different
filenode and point the relation (table or index) at it.

I think the separate space for indexes is a good idea.  However, and
this is orthogonal, I feel the way REINDEX works now is not the best,
because it precludes you from using the index while you are doing it.

I'm trying to implement a way to concurrently compact the indexes.
I hope to have it for 7.4.

-- 
Alvaro Herrera (<alvherre[a]atentus.com>)
Y una voz del caos me hablo y me dijo
"Sonrie y se feliz, podria ser peor".
Y sonrei. Y fui feliz.
Y fue peor.