Hello, Instead of using 3 disks in RAID-0 and one without RAID for archive, I would rather invest into one extra disk and have either a RAID 1+0 setup or use two disks in RAID-1 for the WAL and two disks in RAID-1 for the main database (I'm not sure which perform better between those two solutions).
RAID-1 will give you about twice as fast reads as no RAID (and RAID 1+0 will give you twice as fast as RAID 0), with no significant penalty for writing, and it'll save a lot of manpower in case on disk dies. If you can afford hot-swappable disks, you can even replace a failed disk live, in a few minutes, with no failure at software level. Everything can be remotely setup, including adding/removing a disk from RAID array, if you use Linux software RAID (mdadm), except of course the physical swap of the disk, but that can be done by a non-technician. This solution costs only one extra disk (which is quite cheap nowadays) and will deliver enhanced performances and save a lot of manpower and downtime in case of disk breaking. Regards, -- Gaël Le Mignot - g...@pilotsystems.net Pilot Systems - 9, rue Desargues - 75011 Paris Tel : +33 1 44 53 05 55 - www.pilotsystems.net Gérez vos contacts et vos newsletters : www.cockpit-mailing.com -- Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance
