Summary: sunfire v880 mpxio mirror

From: Jacob Ritorto <>
Date: Mon Sep 10 2007 - 11:22:09 EDT
We've opted to keep it simple and eliminate MPxIO from the solution
because it's not going to help or hurt the project and adds
unnecessary complexity.   So this is a plain old boring S.A.M.E.
across multiple controllers using SVM.  Keep moving.  Nothing to see
here.. :)  SUN call this "split backplane" configuration.

As to the backplane cabling, watch it!  You have to attach a $$$ cable
from the new controller to the expansion backplane -- but when you do,
there are mislabeled connectors on some of the backplanes out there.
Mine was one of them and the disks failed to show up because the
connections were mislabelled.  When I reversed the cables, the disks
showed up during probe-scsi-all.  See documents 806-6592-11,
817-4411-10 and 806-6597-11 on sunsolve for details.  Unfortunately,
I'm unable to re-find the errata document that details the mislabeled
connector issue..  But it's out there on sunsolve.

	Thanks to Darren Dunham, Rainer Heilke and Brad Morrison for the input.

Jacob Ritorto

On 8/9/07, Jacob Ritorto <> wrote:
> Greetings,
> Scenario:
>       We have a v880 with expansion FCAL backplane and additional FCAL
> controller (qlc).  We wish to mirror the original set of disks on the
> original backplane to the new set of disks on the expansion backplane.
>  I'm handy with SVM, but having MPxIO in the mix seems to add a layer
> of complexity (and fault tolerance) I wasn't prepared for.
> Questions:
>       Is this a reasonable way to increase storage fault tolerance?
>       Should I use MPxIO or SVM or both?
>       If we go with SVM alone, is it OK to have just a 2 way mirror on
> all volumes?
>       How is it done? :)  Pointers to best practice docs / recipes
> would be greatly appreciated.
>        If I need to scrap this plan completely and start anew, please
> feel welcome to say so.
> thanks, will summarize.
> Jacob Ritorto
sunmanagers mailing list
Received on Mon Sep 10 11:22:28 2007

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:44:06 EST