On 28/01/16 03:34 AM, Christian Brunotte wrote: > An official "Debian Rescue CD" would seem an unnecessary waste of > storage and buildd resources if it were 99% identical with the "Live > CD". If it contained the Live CD and additional tools, on the other > hand, who would still be downloading the inferior Live CD any longer :-)
It would not be 99% identical. The rescue CD prior to its removal contained quite a number of tools. Also, it did not contain a desktop. So there was very little overlap with the other flavours (except for the base packages). Go look up the old wheezy images to see what I mean. The purpose of the standard set of images, I'll reiterate, is to provide what you would normally end up with after a regular Debian install. As such, that set contains: standard (no desktop), gnome, kde, xfce, lxde, cinnamon and mate images, i.e. all of the desktop choices that you would have with a netinst. While other flavours beyond these are certainly possible, in each case, the list of packages that goes into these are the responsibility of other teams, who have particular interest and skills in maintaining those flavours. We are not that sort of team. > Would it really be so difficult to pass a list of package names to the > installer or at least add the .deb to the CD It is not "the CD", by the way, but a whole set of images. > and leave it to the user to > actally install them with apt-get? Would it help if we had a "task-rescue" > meta package? The technical difficulty of containing a single extra package is not the blocker for this wishlist. Responding to wishlists for the inclusion of extra packages is simply contrary to the goals and vision of this project. Once you respond to one request to include a special extra package, it is a wedge that opens the door for all sorts of other special requests for "useful" things to be added, each with unforeseen consequences to add them. Before long, you no longer have the default Debian install experience, but a whole mishmash of different extra things, at additional cost to the live team, and Debian infrastructure. > The problem with all the other rescue cd projects is that some of them become > abandoned or get only updated every year and do not feature kernels that are > recent enough to boot on modern servers. E.g. I couldn't get a Dell server to > boot with Grml or Gparted-Live despite playing around with "noacpi" or similar > kernel options. I think you argue my point pretty well. As a blend project, it would guarantee there is a team to keep the images updated. As a list within live-images, prior to its removal, 'rescue' flavour was not getting the attention it needed, as this is not the live team's area of focus. > Debian's Live-CD is based on all the work that has gone into > the regular Debian distro to get it booted anywhere. Right. So additionally, a blend would add the value of an expertly chosen set of rescue packages, representing the best of what Debian has to offer in helping a user deal with a wide variety of rescue scenarios. There seem to be a number of people responding to this bug who are willing to help with this, so I consider the issue of fulfilling the original request of this bug resolved. Ben
