Hey Evan, > Sarah Jelinek wrote: >> Hi Evan, >> >>> Hi Sarah, >>> >>> I'm a bit slow getting back to you... >>> >> No worries.. I am slow getting back to you too :-). >> >>> 1.4 and 1.7 isn't a BE just another version of an installed >>> instance? I was thinking that 1.7 should be under 1.4 and that we >>> may want to point out that there is a requirement on beadm/libbe for >>> validaty chekcing for the BE. >>> >> hmm... that's interesting. Well, the way I was thinking about this is >> that we could have an empty BE. Without a zpool or an instance of >> OpenSolaris, right? Is this possible? Is it worth finding? > > While it's possible to create an empty set of datasets that can be > installed into (using be_init in libbe) it doesn't seem that it's a > valid BE until it has bits in it. Currently there is no way to have an > empty BE without at least an existing root zpool. > > I guess since it is possible to create an empty set of datasets that > correspond to a BE we could think of it as something different. What I > was thinking is that beadm/libbe should also be doing at least some > validation of BE's to insure that they are real BE's. While we don't > do that now we should only be thinking of BE's in terms of having > valid bits in them.
Ok, good point. I think I will move 1.7 as an instance of OpenSolaris in 1.4. > >> >> >>> 1.5.c "zpool datasets" should probably be "zfs filesystems on the >>> zpool." >>> >> I put datasets since they are called datasets. To distinguish them >> from the filesystems, such as UFS. > > Ok I that makes since. Maybe use ZFS datasets instead of zpool datasets? Sure, I will change this. > >> >>> 1.6 may be the same kind of thing as 1.7? >>> >> How are they the same? Can't we have zones in a BE? So, would we want >> to find what zones their might be? >> > > I was thinking more that a zone was another form of installed instance > of OpenSolaris but yes a BE can have multiple non-global zones in it. > Based on this yes we would want to find all the BE's, all the > non-global zones within each BE and all the zones BE's for each zone. > > I guess these are all installed instances however they are very > different in how they are accessed so what you had may be more correct > then what I was thinking. I think at this point I will leave these two as separate requirements. thanks, sarah ***** > > > -evan > >> >> thanks, >> sarah >> **** >>> 2.9 Ethan already mentioned this one... >>> >>> thanks, >>> -evan >>> >>> Sarah Jelinek wrote: >>>> Please review and provide comments by COB Wed, 6/17. >>>> >>>> thanks, >>>> sarah >>>> **** >>>> >>>>> Hi All, >>>>> >>>>> I have posted a set of requirements defined for the Caiman unified >>>>> engine. They are located at: >>>>> >>>>> http://opensolaris.org/os/project/caiman/CUD/cud_req.txt >>>>> >>>>> There are some more details that need to be added, specifically >>>>> with regard to observability requirements. A lot of that depends >>>>> on the consumers of the unified engine. Along with the error >>>>> handling section. >>>>> >>>>> Please review and send comments. >>>>> >>>>> Regards, >>>>> sarah >>>>> >>>>> >>>>> _______________________________________________ >>>>> caiman-discuss mailing list >>>>> caiman-discuss at opensolaris.org >>>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >>>> >>>> _______________________________________________ >>>> caiman-discuss mailing list >>>> caiman-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >>> >> >
