On Tue, Apr 07, 2009 at 12:12:12PM -0400, Dave Miner wrote: > The change I had suggested was to use the standard paths and names, > rather than ones which were specific to the live CD (especially as they > were initially architecture-specific). That didn't get translated > correctly in the prior iteration.
I see - that makes sense now, thanks. > >We'll work to get this change out to Linux, EC2 and any other > >stakeholders as soon as we can. I'll repurpose 6826097 to this effect. > >We have: > > > >/boot/x86.microroot (2009.06 and earlier live CD) > >/platform/i86pc/boot_archive (later live CD releases) > >/boot/boot_archive (pre-release 2009.06 bits) > > > >What will the AI release do? Can we drop the third check, or will it > >stay at /boot/boot_archive? > > > > If it would be helpful in reducing the clutter, it looks like it would > be easy to revert it to /boot/x86.microroot for now. We can just > pretend a few builds didn't exist ;-) Just so we're clear, virt-install will look for /platform/i86pc/boot_archive, then /boot/x86.microroot, and then give up. This will cover both AI and live CDs that have been released, and break a few pre-111 iterations. You (or someone) will take care of placing the boot archive in x86.microroot for both live and AI in 2009.06. The next release will modify both to use /platform/i86pc/boot_archive Is that accurate? I should be able to get a 111-based virt-install change done today that does the above. When AI support comes in for virt-install we'll make sure it works similarly. > >>4. Install QE and xVM QE devise a plan for regular regression testing of > >> PV OpenSolaris live CD installs > > > >I'm not sure who's owning this. I'll make sure it's covered again. > > Wasn't meaning to dump that on you necessarily, but if you're game for > following up, that's fine. I'm following up. thanks, john
