[BUGS] BUG #6034: pg_upgrade fails when it should not.

2011-05-22 Thread Tim Uckun
The following bug has been logged online: Bug reference: 6034 Logged by: Tim Uckun Email address: timuc...@gmail.com PostgreSQL version: 8.4 Operating system: ubuntu linux Description:pg_upgrade fails when it should not. Details: pg_upgrade from 8.4 to 9.0 fails

Re: [BUGS] Problems with adding a is not null to a query.

2011-01-17 Thread Tim Uckun
> > As I've stated repeatedly, your next move needs to be to increase the > stats target, at least for that column if not globally.  You probably > don't need to have it know about every last domain id, but you need to > have it know about enough that it realizes that domains not included in > the

Re: [BUGS] Problems with adding a is not null to a query.

2011-01-17 Thread Tim Uckun
> With a table that large, you're probably going to need a larger stats > target in order to get reasonable estimates for low-frequency values. > Am I right in guessing that pg_stats.n_distinct is much too low for > the domain_id column? the domain_id is in the topical urls. A select count of doma

Re: [BUGS] Problems with adding a is not null to a query.

2011-01-17 Thread Tim Uckun
> > Hmm.  What do you get for: > > SELECT relname, pg_relation_size(oid), reltuples, relpages FROM > pg_class WHERE relname IN ('consolidated_urls', > 'consolidated_urls_pkey'); > relname | pg_relation_size | reltuples | relpages +--+-

Re: [BUGS] Problems with adding a is not null to a query.

2011-01-16 Thread Tim Uckun
> > Hmm, autovacuum *should* have been keeping track of things for you, > but it might still be worth doing a manual ANALYZE against that table > to see if the estimated rowcount changes.  If not, you'll need to raise > the statistics target for that column (and again ANALYZE). The analyze finish

Re: [BUGS] Problems with adding a is not null to a query.

2011-01-16 Thread Tim Uckun
> Hmm, autovacuum *should* have been keeping track of things for you, > but it might still be worth doing a manual ANALYZE against that table > to see if the estimated rowcount changes.  If not, you'll need to raise > the statistics target for that column (and again ANALYZE). > I started a manual

Re: [BUGS] Problems with adding a is not null to a query.

2011-01-15 Thread Tim Uckun
> > Possibly the table's never been ANALYZEd ... do you have autovacuum > enabled?  If it has been analyzed reasonably recently, then it might be > necessary to crank up the statistics target to get a better estimate. > It's difficult to give detailed advice when you haven't mentioned what > PG ver

[BUGS] Problems with adding a is not null to a query.

2011-01-15 Thread Tim Uckun
I reported this in the pgsql-general list and was instructed to send the analaze outputs here. have this query it runs reasonably quickly (but should be quicker IMHO) SELECT "consolidated_urls".* FROM "consolidated_urls" INNER JOIN "topical_urls" ON "consolidated_urls".id = "topical_urls".conso