On Mon, Jan 11, 2010 at 3:53 AM, Anton Belyaev <anton.bely...@gmail.com> wrote: > 2010/1/9 Scott Marlowe <scott.marl...@gmail.com>: >> On Fri, Jan 8, 2010 at 3:07 PM, Greg Smith <g...@2ndquadrant.com> wrote: >>> Basically, you have a couple of standard issues here: >>> >>> 1) You're using RAID-5, which is not known for good write performance. Are >>> you sure the disk array performs well on writes? And if you didn't >>> benchmark it, you can't be sure. >> >> This can be doubly bad if he's now moved to a set of disks that are >> properly obeying fsync but was on disks that were lying about it >> before. >> > > Scott, thanks for the interesting suggestion. > > And how do I check this? > Old RAID-1 has "hardware" LSI controller. > I still have access to old server.
The old RAID card likely had a battery backed cache, which would make the fsyncs much faster, as long as you hadn't run out of cache. While RAID-5 can have reasonable throughput sequentially, it has a much higher random write costs, (two reads and two writes for each random write). Likely those two things, lack of cache, and more expensive writes is why the new system seems so much slower. If you can shoehorn one more drive, you could run RAID-10 and get much better performance. -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general