[SUMMARY]: Striped Mirrors

From: Dirk Bönning <dirk.boenning_at_ccf.de>
Date: Mon Nov 18 2002 - 11:33:09 EST

here's the summary:

Original Question:
 > yes, you got the subject right.
 > One of our customers want to know, if it's possible to stripe Mirrors
 > (call it Raid 1+0). I tried the GUI of SDS and it didn't work. Could
 > anybody confirm/defirm that it will work under SDS command line or if
 > this is even possible under Veritas or any other Product ?
 > TIA, Dirk.

Very short version:
-Yes Raid 1+0 exists.
-Yes plain Raid 1+0 is more reliable than plain Raid 0+1
-Yes Veritas do it since 3.11
-Yes SDS do it also (even it's better to call it intelligent Raid 1+0)

Longer version:
I got a big reply and these replies ranges from NO AT ALL upto YES, for 
sure. So I'm not alone in the world of Raid confusion ;-)

The best overview of Raid 0+1 versa Raid 1+0 could be found at:
But please remember they just talking 'bout the plain concepts.

Veritas can do Raid 1+0, but just since Version 3.11. In veritas slang 
they are called StripePro.

Now what's 'bout SDS ?
Most people anwsered that SDS can't handle it. Then I got a link to a 
Sun site which gives the following statement:

 >Solstice DiskSuite software intelligently provides RAID 1 + 0 
 >capabilities when the underlying devices are identical physical disks, 
 >but provides RAID 0 + 1 when, as in the case of network storage 
 >devices, logical devices can support internal redundancy.

Cause not believing anything I read, I start testing myself with SDS, 4 
disks and a long list of different failures scenarios. The result:
Yes, it's true. SDS delivers the reliabilty of Raid 1+0, even if it's 
more an kind of intelligent Raid 0+1. I'm able to expand the Sun 
statement to "identical physical partitions (no need for the whole disk!)".

Hope it helps to minimize the confusion.

CU, Dirk.

Dirk Bvnning
Daimlerstrasse 10
D-64347 Griesheim

t. +49 (0)6155/6668-38
f. +49 (0)6155/6668-68
e. dirk.boenning@ccf.de
w. www.ccf.de
sunmanagers mailing list
Received on Mon Nov 18 11:57:21 2002

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:58 EST