Re: [PERFORM] Slow queries after vacuum analyze

2012-12-21 Thread Kevin Grittner
Ghislain ROUVIGNAC wrote: >> I would leave default_statistics_target alone unless you see a lot of >> estimates which are off by more than an order of magnitude. Even then, it >> is often better to set a higher value for a few individual columns than for >> everything. > > > We had an issue with

Re: [PERFORM] Performance on Bulk Insert to Partitioned Table

2012-12-21 Thread Charles Gomes
The BBU does combines the writes. I've benchmarked using a single table and it took 1:34:21.549959 to insert 118800 rows. (70 writers to a single table) I've also benchmarked having writers targeting individual partitions and they get the same job done in 1 Hour. I/O is definitely not the

Re: [PERFORM] Performance on Bulk Insert to Partitioned Table

2012-12-21 Thread Charles Gomes
Tom, I may have to rethink it, so I'm going to have about 100 Million rows per day (5 days a week) 2 Billion per month. My point on partitioning was to be able to store 6 months of data in a single machine. About 132 partitions in a total of 66 billion rows. ---