[ale] RAID, IDE and/or Linux

David Corbin dcorbin at imperitek.com
Sat Feb 16 19:52:51 EST 2002


I think Jeff's point was that failover *is* instaneous.  But you're 
right that for sometime afterward, performance is "below par".


Chris Ricker wrote:

>On Sat, 16 Feb 2002, Jeff Hubbs wrote:
>
>>Chris Ricker wrote:
>>
>>>On Sat, 16 Feb 2002, Jeff Hubbs wrote:
>>>
>>>>The downside to RAID 1 here is the 1:2 effective/actual capacity ratio, 
>>>>i.e., takes 40GB to make 20.  If he goes RAID 5 instead and the 
>>>>three-drive medium, he gets a ~2:3 ratio - takes 60GB to get 40.  
>>>>
>>>But loses instantaneous fallover....
>>>
>>Since when??
>>
>
>Since always.  By definition, with RAID-5, you don't have a live copy of
>data like with RAID-1 -- you have dispersed parity bits from which data can
>be slowly reconstructed.  Even when you have hot spare(s) (and I hope if
>you're using RAID-5 that you do), you still have to wait for the new drive
>to be rebuilt from the parity on the remaining disks.  Until the new drive
>finishes rebuilding, you run in a degraded mode where data is reconstructed,
>not live.
>
>later,
>chris
>
>
>---
>This message has been sent through the ALE general discussion list.
>See http://www.ale.org/mailing-lists.shtml for more info. Problems should be 
>sent to listmaster at ale dot org.
>
>




---
This message has been sent through the ALE general discussion list.
See http://www.ale.org/mailing-lists.shtml for more info. Problems should be 
sent to listmaster at ale dot org.






More information about the Ale mailing list