Brett Albertson wrote:
> I've been upgrading my Solaris Express OS every time a new Solaris Express 
> came out for the past few months.  The upgrade worked every time until the 
> build where Caiman went in (70 or 69?).  I have my current and active OS 
> installed on c0d1, and an old version of Solaris 10 on c0d0.  The old 
> installer/upgrader always let me choose which to upgrade.  The new installer 
> just goes ahead and upgrades c0d0 without asking me.
> 
> So, I thought I would outsmart it by removing everything in /etc, /kernel, 
> /dev, and /devices on c0d0.  It still tries to upgrade it, but now fails 
> saying it could find /a/etc/vfstab.  I then removed the tag from slice 0 
> called "root" to make it unassigned, and it still tried to upgrade c0d0.  
> 
> So, I have two questions:
> 1. Is this a bug?

Possibly.  It never presents the Disk screen with the to select which 
instance is to be upgraded?  What are the contents of 
/var/sadm/system/admin/CLUSTER on the c1d0 disk?  Section 3.3 of the 
Target Discovery design doc [1] covers the algorithm used in identifying 
installed Solaris instances, so there are a number of things which might 
cause c1d0 to not be considered upgradeable.

Dave

[1] http://www.opensolaris.org/os/project/caiman/files/dwarf_td_design.pdf

Reply via email to