Re: Slow count(*) again...

On 10/09/2010 06:54 PM, Mladen Gogala wrote:
> In another database, whose name I will not mention, there is a parameter
> db_file_multiblock_read_count which specifies how many blocks will be
> read by a single read when doing a full table scan. PostgreSQL is in
> dire need of something similar and it wouldn't even be that hard to
> implement.

You're correct in that it isn't particularly difficult to implement for
sequential scans. But I have done some testing with aggressive read
ahead, and although it is clearly a big win with a single client, the
benefit was less clear as concurrency was increased.

Joe

-- 
Joe Conway
credativ LLC: http://www.credativ.us
Linux, PostgreSQL, and general Open Source
Training, Service, Consulting, & 24x7 Support