At 06:13 PM 2/9/2006 +1100, you wrote:
>Raid-0 is not random, but you are right, there is no redundancy.
It's only random in that it's not predictable by the RAID
controller -- with only one disk having that information on it
you can't use predictive logic to read in advance and are sort
of stuck with read-as-it-comes-around. With disk speeds at
5400, 10K, and 15K rpm this isn't a lot of time we're talking
about, but compared to the access to memory it might as well
be measured in geological terms.
One other thing: SATA vs SCSI. When choosing your server
config SATA and even IDE is a lot cheaper than SCSI. The thing
about SATA and IDE is you get access to one disk at a time from
the controller. SCSI you can access all devices at pretty much
all times. That requires extra intelligence on the SCSI devices
so they tend to be more expensive, but if you want to run a
RAID-5 array on 5 drives you've 4 reads and 5 writes to do for
each bit of data. That's one pipe at one time on SCSI, it's
four independent reads and five independent writes on SATA.
You can see where making this all a one-at-a-time thing can
reduce performance and make SCSI a viable alternative for a
Raid-5 array but probably a waste of money for a raid-1.
Short and sweet, decide what kind of disk IO you're
going to see on your server and choose accordingly. RAID-5
on a CS server is a waste of money and speed, but RAID-5
is probably a good idea on something that you're hosting six
games and a webserver on if you choose SCSI drives.
- Dan
* Dan Sorenson DoD #1066 A.H.M.C. #35 [EMAIL PROTECTED] *
* Vikings? There ain't no vikings here. Just us honest farmers. *
* The town was burning, the villagers were dead. They didn't need *
* those sheep anyway. That's our story and we're sticking to it. *
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds
