Re: [GENERAL] Call for alpha testing: planner statistics revisions - Mailing list pgsql-hackers

From Ed Loehr
Subject Re: [GENERAL] Call for alpha testing: planner statistics revisions
Date
Msg-id 3B324568.9E6CBABE@pobox.com
Whole thread Raw
In response to Call for alpha testing: planner statistics revisions  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane wrote:
> 
> I have finished a first pass at the planner statistics and cost
> estimation changes that I want to do for 7.2.  It's now time to see
> how well the new code does in the real world...
> 
> Some highlights of the new code include:
> 
> * ANALYZE is now available as a separate command; you can run it without
> also doing a VACUUM.  (Of course, VACUUM ANALYZE still works.)

What is the impact of this newly isolated ANALYZE command on the need
and/or frequency for VACUUMs?  

I'd like to reduce the frequency of VACUUMs given my understanding has
been that updates/inserts/deletes are a no-no during VACUUMs (that was
6.5.x era hearsay), and we lock people out during VACUUMs (VACUUM
ANALYZE, that is).  

Regards,
Ed Loehr


pgsql-hackers by date:

Previous
From: Jason Tishler
Date:
Subject: Re: [current] readline breakage
Next
From: Tom Lane
Date:
Subject: Re: SPI_prepare for semi-unknown types