* Eric J. Ray (Eric.Ray at Sun.COM) wrote: > > On Mar 12, 2009, at 10:10 AM, Glenn Lagasse wrote: >>> >>> What's the use case around generating several kinds of images, but >>> only requiring checksums for some of them, or generating different >>> checksums for different kinds of images? Common? >> >> I don't know that there is enough data to determine what's common. If >> we use the OpenSolaris releases as a guide, I'm pretty sure that we >> want >> to generate the same checksums for all the images we create. And so >> requiring checksums for some of them or different checksums for >> different kinds of images is probably not common (at least in the >> OpenSolaris release case). > > So, what does that do for your design?
It makes it simpler. We create a new 'toggle' in the DC manifest to enable checksumming of images (all images) and then the finalizer scripts that generate images compute image hashes using the list of hash types that I deemed appropriate/most useful (md5, sha1, sha256). If there's a need for more flexibility (ability to define what hashes are generated on what images) then that could always be implemented later. -- Glenn
