Tom Horsley wrote: > Mostly that it is really really hard to do right and > you'd probably never get the kind of discipline and > testing require from the wild west open source > community :-). > Maybe it's not all that hopeless. Virtualization progressed in heaps and bounds recently. What can be done is having a simple core kernel doing nothing but managing virtual machines. When it's time to "reboot" we'll start a new virtual machine, and gradually switch the applications to it (can be done automatically or with user assistance, depending on what they do & how they work). Then, after the original VM is not running anything of interest to the user, it's turned off. This adds an advantage of being able to try the new kernel & not switching to it if something doesn't work. Of course it can't be that easy in practice, but it's much easier/safer than doing all the updates in place. -- fedora-list mailing list fedora-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
