For today, at 200GB or less of required space, and 500GB or less next year.
“Where we’re going, we don’t NEED spindles.”
Seriously, go down to the store and get 6 X25-M’s, they’re as cheap as $550 each and will be sub $500 soon. These are more than sufficient for all but heavy write workloads (each can withstand ~600+ TB of writes in a lifetime, and SMART will tell you before they go). 6 in a RAID 10 will give you 750MB/sec read, and equivalent MB/sec in random reads. I’ve tested them. Random writes are very very fast too, faster than any SAS drive.
Put this in your current system, and you won’t need to upgrade the RAM unless you need items in cache to reduce CPU load or need it for the work_mem space.
Spindles will soon be only for capacity and sequential access performance requirements. Solid state will be for IOPS, and I would argue that for most Postgres installations, already is (now that the Intel SSD drive, which does random writes and read/write concurrency well, has arrived — more such next gen drives are on the way).
On 12/9/08 9:28 AM, "Scott Carey" <scott@xxxxxxxxxxxxxxxxx> wrote:
> Lucky you, having needs that are fulfilled by sequential reads. :)
> I wonder how many hard drives it would take to be CPU bound on random
> access patterns? About 40 to 60? And probably 15k / SAS drives to
> boot. Cause that's what we're looking at in the next few years where
> I work.
About $3000 worth of Intel --- mainstream SSD's = 240GB space (6 in raid 10) today, 2x to 3x that storage area in 1 year.
Random reads are even easier, provided you don't need more than 500GB or so.
And with something like ZFS + L2ARC you can back your data with large slow iops disks and have cache access to data without requiring mirrors on the cache ($3k of ssds for that covers 2x the area, then).
--
Sent via pgsql-performance mailing list (pgsql-performance@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
