[Grml] /dev, udev and backup
Michael Prokop
mika at grml.org
Sun Jul 13 01:01:54 CEST 2008
* T o n g <mlist4suntong at yahoo.com> [20080712 07:05]:
> On Fri, 11 Jul 2008 22:15:57 +0200, Michael Prokop wrote:
> > Well, static devices can be found in /dev/.static/dev
> > and if configured/used also in /lib/udev/devices.
> > What's really important are the null, zero, console and tty[0-9]
> > devices which are all easy to generate/restore using MAKEDEV in case
> > you lose them.
> Thanks for the explanation mika. I was hoping that I can just backup those
> static devices, but I found both my /dev/.static/dev/ & /lib/udev/devices/
> are empty. If I duplicate null, zero, console and tty[0-9] devices into
> /dev/.static/dev/ and back them up, would it works?
Just executing MAKEDEV will provide null, zero & CO for you.
No need to copy them to /dev/.static/dev, though I'm wondering why
your /dev/.static/dev/ is empty. Are you listing the content of the
directory as user root ("ls -la /dev/.static/dev/")?
> > So basically it's not necessary to backup /dev as long as you know
> > how to recover it ;) - but as /dev is usually less than 1MB size it
> > won't hurt to place it in your backups as well. ;)
> Yeah, 1MB is not that big. My worry was the portability. If I backup all
> devices in PC A, then restore them on PC B, what would happen?
Relevant are the core devices (as mentioned above) and the device
for your rootfs ("mount | grep 'on / '"), the rest will be handled
by udev and its tmpfs. If you just backup whole /dev you should be
on the safe side.
regards,
-mika-
--
http://grml.org/ # Linux for texttool-users and sysadmins
http://wiki.grml.org/ # share your knowledge
http://grml.supersized.org/ # the grml development weblog
#grml @ irc.freenode.org # meet us on irc
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://ml.grml.org/pipermail/grml/attachments/20080713/0bed0cbe/attachment-0003.pgp>
More information about the Grml
mailing list