Re: Partitioning Vs. Split Databases - performance?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 12/21/06 23:44, Joshua D. Drake wrote:
>>>> With One Big Database, you can get a SAN and attach a whole lot of
>>>> disk space, but your mobo will only accept a certain number of DIMMs
>>>> and processors of certain designs. And when your growing mega
>>>> database maxes out your h/w, you're stuck.
>>> Define mega... Because you would need to be in the multi-terrabyte
>>> range.
>> I'm thinking more of RAM and CPU.
>
> 32GB is an awful lot of ram... as is 8 cores. You can get 16 core
> machines now that will take 64GB.

Sure, you *can*. A Sun Fire X4600 (max 16 cores) takes up to 128GB
RAM. I'll bet my house, though, that OP doesn't have a maxed-out
X4600. (BTW, maxed out, with 2 U320 adapters and 6 4Gb HBAs, it's
only $75000!)

Please tell me, though, what's so controversial about saying that
CPU and RAM resources are finite?

Besides, since pg_dump is single-threaded, backing up a huge
database gets impossible. Federating the database allows multiple
pg_dumps to simultaneously dump data to multiple tape drives.

- --
Ron Johnson, Jr.
Jefferson LA USA

Is "common sense" really valid?
For example, it is "common sense" to white-power racists that
whites are superior to blacks, and that those with brown skins
are mud people.
However, that "common sense" is obviously wrong.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFFi3zlS9HxQb37XmcRAtVjAKDGPu4jHPyTH9EX5sqPbuHs+wKE3gCeLnnx
RvSfELtJ7bieg3HVVqB/7Zk=
=FQCv
-----END PGP SIGNATURE-----