Unfortunately I neglected to cc caiman-discuss... (Thanks Karen for catching this oversight)
I'm working on the draft proposal now and will add any comments to it after I have the draft finished so that people have some time to look over the information for the proposal. I would like to get any comments on this by close of business Tuesday. The link to the proposal information and notes is at: http://opensolaris.org/os/project/caiman/CVERS/proposal/ Thanks, -evan -------- Original Message -------- Subject: Re: Version compatibility meeting. Date: Thu, 28 May 2009 16:24:32 -0600 From: Evan Layton <[email protected]> To: Ethan Quach <Ethan.Quach at Sun.COM>, Karen Tung <Karen.Tung at Sun.COM>, Sarah Jelinek <Sarah.Jelinek at Sun.COM>, Jack Schwartz <Jack.A.Schwartz at Sun.COM>, Clay Baenziger <clayb at sun.com> References: <4A1C523F.9020008 at sun.com> <4A1DE8DF.8090009 at sun.com> I've updated the wiki (http://opensolaris.org/os/project/caiman/CVERS/) with the Notes from todays and yesterday's meetings. For the requirements and use case/rule-set stuff, if possible I'd like to get your comments by about 3pm PT tomorrow afternoon. This is so I can add them to the draft proposal and get that out some time Friday night. (http://opensolaris.org/os/project/caiman/CVERS/proposal/) I know we all have a lot going on right now so if you can't make that time frame please let me know. Thanks! -evan Evan Layton wrote: > We're having another meeting tomorrow to finish going through the items > on the agenda that we didn't get to. That meeting will be at 10am PT > (11am MT) tomorrow morning. We'll be using the same conference call > information. > > Thanks, > > -evan > > > Evan Layton wrote: >> I'd like to have a meeting Wednesday May 27th (tomorrow) at 2pm PT >> (3pm MT). >> >> Agenda: >> - Status of research tasks. >> - Ideas from Danek on how to version IPS bits. >> - requirements >> - Uses cases/rule sets >> - what's in a proposal and would it smell as ... >> >> Call info: >> USA Toll-Free: 866-545-5227 >> USA Caller Paid/International Toll: 215-446-3648 ACCESS >> CODE: 1337371 >> >> >> Thanks! >> -evan > > > Notes From the today's meeting: > > Agenda: > - Status of research tasks. > - Done... > - Target instantiation and target discovery, dependencies on > libdiskmgt. Any new features added may not be available on > installer. This is the same as the current installer. This does > not cause the installer to fail and results will continue to > be consistent it just won't use the new features. > - Shouldn't have boundary issue with libzfs and other libraries > because we have tied the installer to the boot image. > - These things will be less of a problem once we have arc'd > things and have contracts with these groups. > - Need to better understand the possibility for installing older > bits from a newer installer. This would only be available for > a image install (replicated image). > - You install what you booted but we do what we can to make it > match what you expected you are installing. There is a > potential that the install done with an installer that is older > than the bits being installed it is possible that there may > be side effects that can cause the installed system to have > some differences from an install of the same bits but using > an installer booted to the same OS version as the bits being > installed. > - What ICT's are there that have to be done after an image-update > (such as installgrub or bootadm update-archive). > - provide an ict package that we download to give us the icts > for the bits being installed. However these new ict's will be > run on a booted version that may be older than what these > ict's have been tested to work on. > > - Ideas from Danek on how to version IPS bits. > - need to make sure we have the ability search the repo remotely > for the IPS version. > - This may help us with zpool version stuff going backward since > we could get the correct zpool version of the bits being > installed by querying the IPS repo. > - Need to add dependency on IPS allowing us to image-update to > a specific version. This is the main hole in the current design > ideas. > > - requirements > Still needs discussion > > > - Uses cases/rule sets > Still needs discussion > > > > - what's in a proposal > Still needs discussion > >
