Re: [PERFORM] REINDEX takes half a day (and still not complete!)

2011-05-06 Thread Merlin Moncure
On Sat, Apr 30, 2011 at 4:26 AM, Phoenix Kiula wrote: > I suppose that's what I am going to do on a periodic basis from now > on. There is a lot of DELETE/UPDATE activity. But I wonder if the > vacuum stuff really should do something that's similar in function? > What do the high-end enterprise fo

Re: [PERFORM] VX_CONCURRENT flag on vxfs( 5.1 or later) for performance for postgresql?

2011-05-06 Thread Dimitri
You should rather consider VxFS tuning - it has an auto-discovery for DIRECT I/O according the block size. Just change this setting to 8K or 16-32K depending on your workload - then all I/O operations with a bigger block size will be executed in DIRECT mode and bypass FS cache (which logical as usu