Re: [PERFORM] [BUGS] BUG #1552: massive performance hit between 7.4

2005-03-27 Thread Karim Nassar
On Fri, 2005-03-25 at 10:18 +, Simon Riggs wrote: When running with PostgreSQL 7.4 on a dual-CPU Athlon MP2400+ machine with a gigabyte of RAM, running Debian Linux version 2.6.8.1, we were able to insert all this data in 5-7 minutes. It's taken a while to install Postgres 8.0.1 on

Re: [PERFORM] [BUGS] BUG #1552: massive performance hit between 7.4 and 8.0.1

2005-03-25 Thread Simon Riggs
On Wed, 2005-03-23 at 14:22 -0500, Keith Browne wrote: Simon Riggs wrote: The EXPLAINs you've enclosed are for SELECTs, yet your bug report describes INSERTs as being the things that are slow. [You may find better performance from using COPY] We're starting with an empty database,

Re: [PERFORM] [BUGS] BUG #1552: massive performance hit between 7.4 and 8.0.1

2005-03-25 Thread Tom Lane
Simon Riggs [EMAIL PROTECTED] writes: I think we should spawn a TODO item from this: * Coerce FK lookups to always use an available index No, we aren't doing that. The correct TODO item is Replan cached plans when table size has changed a lot which of course depends on having a framework to do