[rear-users] Restoring mdadm system to a single-disk destination fails
dag at wieers.com
Tue Jul 24 23:34:15 CEST 2012
On Tue, 24 Jul 2012, Les Mikesell wrote:
> On Tue, Jul 24, 2012 at 10:39 AM, Dag Wieers <dag at wieers.com> wrote:
>> What we currently do is IMO
>> the best we can do (minus the resizing of the swap). Maybe there is a case
>> to influence the resizing of the swap (or maybe even go swapless), but
>> since this is already possible by modifying the disklayout.conf, I am not
>> convinced that having a switch to decide this in advance is very useful.
> Maybe things have changed since the last time I converted an MD system
> to /dev/sd? partitions and had to rebuild the initrd by hand but I
> would have appreciated:
> 1) numbers in disklayout conf that matched some other tool I might
> have used before and/or a tool to help match them to my new available
> target device(s). Computers are good at that kind of math - I'm
There is a good reason we use bytes, it's because we have experience with
various versions of parted, of which some cannot do anything other than
bytes. So bytes always works.
I don't mind if we would add the capabilities to check for such a feature
and use something more human readable in case parted supports it. But this
might also affect the other tools.
> 2) a reminder that /etc/fstab and grub.conf needed to be fixed
As I already explained, we already fix /etc/fstab and grub.conf. It may
not be sufficient, but that we don't know. And if we know, we might as
well do it ourselves... What is needed is a detailed bug-report when it
doesn't work, a description on how to reproduce it or someone contributing
this piece of code. (Making an issue out of it is a good start, but only
do so if you have enough details for someone to dive into it directly)
> 3) a scripted option to chroot to the mounted system (also available
> after a reboot from the iso without repeating the restore or manually
> mounting the partitions)
Please add to the issue-tracker together with scenarios where this could
be useful ! Ideas that build on this are worth bonus points...
> 4) any automated help possible with making a new initrd. This
> happened to be under vmware where the boot iso had a working driver
> but the restored system did not, but there are even worse
> possibilities. Ideally the boot iso would work like the distribution
> installer and be as general as possible, then help you set up the
> modules needed for the devices detected - but that might be too much
> to ask...
Please report your issue in the issue-tracker with as much detail as
possible to reproduce.
-- dag wieers, dag at wieers.com, http://dag.wieers.com/
-- dagit linux solutions, info at dagit.net, http://dagit.net/
[Any errors in spelling, tact or fact are transmission errors]
More information about the rear-users