On Mon, 2006-08-21 at 15:45, Marty Jia wrote:
> I'm exhausted to try all performance tuning ideas, like following
> parameters
>
> shared_buffers
> fsync
> max_fsm_pages
> max_connections
> shared_buffers
> work_mem
> max_fsm_pages
> effective_cache_size
> random_page_cost
>
> I believe all above have right size and values, but I just can not get
> higher tps more than 300 testd by pgbench
>
> Here is our hardware
>
>
> Dual Intel Xeon 2.8GHz
> 6GB RAM
> Linux 2.4 kernel
> RedHat Enterprise Linux AS 3
> 200GB for PGDATA on 3Par, ext3
> 50GB for WAL on 3Par, ext3
>
> With PostgreSql 8.1.4
I assume this is on a blade server then? Just guessing. I'd suspect
your vscsi drivers if that's the case. Look into getting the latest
drivers for your hardware platform and your scsi/vscsi etc... drivers.
If you're connecting through a fibrechannel card make sure you've got
the latest drivers for that as well.
1500, btw, is quite high. Most fast machines I've dealt with were
hitting 600 to 800 tps on fairly good sized RAID arrays.
You may be able to put your pg_xlog on a sep partition / set of spindles
and get some perf gain. Also look into how your drives are configured.
The more drives you can throw into a RAID 10 the better. RAID 5 will
usually never give as good of write performance as RAID 10, although it
gets better as the number of drives increases.