[ale] grub stupid question

[Jim Kinney]

Did not mean to take it offline.
On Jan 31, 2015 11:21 AM, "Michael B. Trausch" <mike at trausch.us> wrote:

>  On 01/31/2015 09:38 AM, Jim Kinney wrote:
>
> Containers as in docker/containers or something else?
> On Jan 31, 2015 9:37 AM, "Michael Trausch" <mike at trausch.us> wrote:
>
>>  I just use QCOW2 containers. When things get borked, it's very easy and
>> efficient to restore. :)
>>
>
> (BTW - did you mean to take this off-list?)
>
> QCOW2 is a container file format for virtual block devices, allowing
> snapshots, etc.
>
> A couple months back I had to do a complex staged install of a system that
> I didn't have good accessibility for.  So I used QEMU to install the OS,
> staged on two QCOW2 HDDs and then I took the QCOW2 images and wrote a small
> statically-linked utility to write them from a special-purpose initramfs.
>
> It turns out, QCOW2 makes for a *VERY* convenient format for block image
> backup in its own right, without the use of VM software.  Added bonus is
> that it can be dumped straignt into a VM in order to do testing/debugging
> on it.
>
> Restoration from a saved QCOW2 image is very simple: load the L1 table,
> load the L2 tables, walk them to find the actual used clusters, and read
> the stored clusters and write them to the correct location on the target
> device.  Going the other way is pretty simple, too: Read clusters from HDD,
> if they aren't all-zeroes, write them to the QCOW2 image.
>
> It's also a great, simple format for using to work around Windows
> brain-deadness.  For example: working with common Rπ SD images from Windows
> when actually on an SD card?  Impossible.  (Well, not completely. Just
> impossible in the default configuration of a Windows system.)
>
> So I'm talking about a very low-level use of QCOW2, for a purpose other
> than its intended, albeit one for which it serves quite nicely.
>
>     — Mike
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ale.org/pipermail/ale/attachments/20150131/2b12ac10/attachment.html>