| From: | Gregory Stark <stark(at)enterprisedb(dot)com> |
|---|---|
| To: | "Scott Marlowe" <scott(dot)marlowe(at)gmail(dot)com> |
| Cc: | <robert(at)webtent(dot)com>, "Bill Moran" <wmoran(at)potentialtech(dot)com>, "PostgreSQL" <pgsql-general(at)postgresql(dot)org> |
| Subject: | Re: autovacuum |
| Date: | 2007-09-20 22:08:27 |
| Message-ID: | 87vea59ehw.fsf@oxford.xeocode.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
"Scott Marlowe" <scott(dot)marlowe(at)gmail(dot)com> writes:
> RAID5 optimizes for space, not performance or reliability. It gets
> faster but less reliable as it gets bigger. If you can afford the
> space RAID-10 is generally preferred.
RAID5 can be faster for DSS style work loads. If you're writing data to the
raid in large contiguous chunks then it you get higher bandwidth than RAID1+0.
The problem with RAID5 is that if you're writing random access chunks then
it's even slower than not having a raid at all.
--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Merlin Moncure | 2007-09-20 22:38:41 | Re: Migration from PervasiveSQL |
| Previous Message | Tom Lane | 2007-09-20 22:05:38 | Re: Adding domain type with CHECK constraints slow on large table |