Ethan Quach wrote: > > > Tim Knitter wrote: >> >> >> Ethan Quach wrote: >>> >>> >>> Tim Knitter wrote: >>>>> If it didn't exist, or if it failed for whatever reason, I think we >>>>> should >>>>> err on the global side and assume we're in the global rather then a >>>>> non-global shouldn't we? >>>>> >>>> >>>> If it doesn't exist an OSError is raised and execution halts. If it >>>> returns garbage then there is probably something inherently wrong >>>> and allowing beadm to continue probably isn't correct either. I >>>> don't think we should assume either and halt execution in either case. >>>> >>>> >>> >>> The issue I see with this is that beadm doesn't inherently depend on >>> zonename for it to run properly in the global zone, so why make it? >>> >> >> True, however I sure don't feel comfortable allowing the code to >> continue if there is OSError raised. Then the check to see if we're in >> a NGZ will be suspect at best. Are you suggesting another check if >> zonename fails or just continuing? > > Would this be any worse than it is today? I'm suggesting to just continue. >
No. OK I continue now instead of hard fail. webrev updated. Thanks Tim > -ethan > > >> >> Tim >> >>> So I'd rather it be that you have to prove you're in a non-global zone >>> for beadm not to run. >>> >>> >>> -ethan >>>
