Re: Why compiling.
- To: email@example.com
- Cc: debian-user <firstname.lastname@example.org>
- Subject: Re: Why compiling.
- From: Celejar <email@example.com>
- Date: Tue, 31 Jul 2012 13:29:51 -0400
- Message-id: <firstname.lastname@example.org>
- In-reply-to: <4FFD79EC.email@example.com>
- References: <CAGWVfM=vWd6HVuuc2WYFSiQZexqr3NL9R7CnSH2+r_bj6MUryA@mail.gmail.com> <4FFC344B.firstname.lastname@example.org> <email@example.com> <20120710202246.GO5583@sid.nuvreauspam> <4FFCB8A5.firstname.lastname@example.org> <email@example.com> <4FFCF141.firstname.lastname@example.org> <email@example.com> <4FFD79EC.firstname.lastname@example.org>
On Wed, 11 Jul 2012 09:04:44 -0400
Gary Dale <email@example.com> wrote:
> On 10/07/12 11:28 PM, Celejar wrote:
> > On Tue, 10 Jul 2012 23:21:37 -0400
> > Gary Dale<firstname.lastname@example.org> wrote:
> >> On 10/07/12 10:52 PM, Celejar wrote:
> >>> On Tue, 10 Jul 2012 19:20:05 -0400
> >>> Gary Dale<email@example.com> wrote:
> > ...
> >>>> Having a portable kernel is a lot simpler than trying to rescue a
> >>>> non-bootable machine from a live CD.
> >>> True - but then I can just grab a distro stock kernel before I swap
> >>> HDDs.
> >> You still need to go through the aggravation of booting from a live CD
> >> then setting up a chroot environment just to get around the fact that
> >> you compiled a non-portable kernel. You wouldn't have to do any of that
> >> if you had just stuck with the stock kernel.
> > I must have misunderstood what you meant. If machine A is non-bootable,
> > then I need to recover using resources from machine B. But even if
> > machine B generally runs my custom kernel, before I pull its HDD and
> > move it to A, I can just add a stock kernel to B. Can you explain what
> > you mean here?
> The reason machine A is not bootable is because a minor hardware change
> is capable of doing that with a custom kernel. I'm not talking about
> using a machine with a custom kernel to rescue another machine, I'm
> talking about the much greater chance that a machine with a custom
> kernel will need rescuing.
Oh, okay. But this is just an argument for keeping at least one generic
kernel on the machine, which I always do.