Page 1 of 1
Only ext3 ?
Posted: Fri Nov 28, 2008 2:34 pm
do fsarchiver works only with ext3 or with any from the linux kernel supportet file system at this moment?
is it possible to backup the root device as ext3 and restore it as reiserfs and it is still bootable?
Posted: Fri Nov 28, 2008 6:04 pm
At the moment (version 0.1.3) it can only save the filesystem attributes for ext2/ext3. I mean it will preserve the filesystem label, uuid and block size if you save an ext2 / ext3 partition. I will implement the same thing for other filesystem as soon as possible.
Yes it should still be bootable as long as you make sure grub is still ok (you may have to rerun grub-install), and if you update your /etc/fstab. You may also have to update grub configuration if you define the root filesystem using something like "root=LABEL=/" or "root=UUID=xxx" instead of /dev/xxx.
Anyway there may still be bugs at the moment, so just use it on a test environment.
Posted: Sat Nov 29, 2008 11:23 pm
How is the status for xfs?
When backup xfs and restore it to the same partitions,
would it work?
Posted: Sun Nov 30, 2008 3:52 pm
The current version (0.1.5) is able to save/restore all linux filesystems (ext2, ext3, ext4, reiserfs, reiser4, xfs, jfs, btrfs), but at the moment only ext2/ext3 and reiserfs have a specific filesystem attributes support.
It means if you save an ext3fs, your filesystem label/uuid will be preserved when you restore, but if you use an xfs, the filesystem attributes won't be preserved. The contents of the filesystem (files & directories) must be ok on an xfs anyway, including the file attributes (permissions, atime/mtime, ...)
I am focused on implementing a good support for ext2/ext3/ext4 and reiserfs first, since these filesystems are widely used, and I will work on the filesystem attributes for xfs, jfs, reiser4, btrfs as soon as I have time to do it.
Posted: Mon Dec 01, 2008 11:53 am
an interesting project.
when i using a newer version of ubuntu, this works with the uuid in the fstab. (i ever know, that this uuid would be a problem).
when i use the stable version of debian 4, the fstab contains still the old way, /dev/sdax and not the new uuid. in this way, fsarchiver should work with any filesystem like xfs or reiserfs. correct?
Posted: Mon Dec 01, 2008 7:37 pm
Yes the physical name (/dev/sdaX) will always work. When the name is relative to a name written in the superblock of the partition (label or uuid) then it's important that fsarchiver preserves these attributes. In the current version it's only preserved for ext2/ext3 and reiserfs but I will work on supporting that for all filesystems.
Posted: Tue Dec 02, 2008 11:49 am
yes in this way i think.
is the uuid an random number, so that every uuid is different,
or is there a way to gibe the mkfs an uuid for an new partition to create?
why is the uuid used? i see no good thinks in it.
on debian i had big problems. debian 4 use the /dev/sdax way, and the new debian 5 uses the uuid. now i try to upgrade from debian4 to debian5 with aptitude, it works fine. but afer rebooting the new kernel crashes with "no root device found". i think, that in the fstab there are wrong entrys. in this way i see do advantage
Posted: Wed Dec 03, 2008 7:42 am
An UUID is a random number created when you run mkfs. It allows you to be sure that each partition has a unique identifier.
You can use labels or uuid to identify partitions when you boot, using options such as "root=LABEL=/" or "root=UUID=xxx". That way you system can still boot even when the root filesystem is on a different partition / hard-disk.
Posted: Wed Jan 14, 2009 11:57 am
how is the support for ext4 at the moment.
in the future, ubuntu 9.04 will have ext4 support, so that i think about using ext4 at the boot partition. with partimage a backup of ext4 is not further possible. so i must work with fsarchiver.....
Posted: Wed Jan 14, 2009 1:07 pm
ext2, ext3, ext4 are just the same filesystem with different attributes. ext4 should work but it has not be tested. I only made tests with ext3 for now, I will be happy to know if it works for you.