Re: [HACKERS] [hackers]development suggestion needed
От | Don Baccus |
---|---|
Тема | Re: [HACKERS] [hackers]development suggestion needed |
Дата | |
Msg-id | 3.0.1.32.20000114081444.010896f0@mail.pacifier.com обсуждение исходный текст |
Ответ на | Re: [HACKERS] [hackers]development suggestion needed (Adriaan Joubert <a.joubert@albourne.com>) |
Список | pgsql-hackers |
At 07:58 AM 1/14/00 +0000, Adriaan Joubert wrote: >>You don't get to do either of the latter two unless you write a > >> >raw-device storage manager >> >> Not within a single filesystem, but scattering things across spindles >> could be done without a raw-device storage manager :) > >Yes, but seen how cheap RAID arrays have become? I know disks are getting >bigger as well, and many people will opt for a single disk, but there may >be more urgent things to fix than something for which a hardware solution >already exists. And lets face it: a database ought to be on RAID >anyway,unless somebody wants to write Tandem-style mirrored disks.... ;-) Don't need to write Tandem-style mirrored disks, the Linux kernal implements mirrored file systems for me. I can mirror UW2 disks in software for $189/spindle (current cost of an IBM Deskstar UW2 7200 RPM 4.5 GB spindle here in Oregon), the fancier RAID arrays still aren't that cheap. The cheapest RAID interfaces just hide the mirroring from you. There's a tier up that take a cluster of mirrored (or RAID 5'd) platters and present them to you as a single large disk - these remove a lot of one's control over spindle placement, sure. My guess is that some folks don't view this as a plus... - Don Baccus, Portland OR <dhogaza@pacifier.com> Nature photos, on-line guides, Pacific Northwest Rare Bird Alert Serviceand other goodies at http://donb.photo.net.
В списке pgsql-hackers по дате отправления: