I'm afraid that I don't really have any GNU/Linux experience in this
area. But I do know a bit about it from HP/UX and know that the LVM for
GNU/Linux is very similar. If you have LVM(2?) available to your chosen
distro, you can use lvcreate -s to initialise a snapshot logical volume
which is a frozen version of the original. The original lv can continue
to be updated while you backup the snapshot, or whatever. According the
man page of lvcreate (from Fedora Core 4), you should have 15-20% of the
size of the snapshot LV available to create the extra LV.
The LVM HOWTO (http://www.tldp.org/HOWTO/LVM-HOWTO/index.html) has some
info in section 3.8
I don't know what state files are left in in the snapshot image. What I
mean by that is, I don't know how valid/safe it is to have a database
running on a LV, snapshot it, backup the snapshot and then remove the
snapshot.
In the HP/UX scenraios that I looked at, the preferred method was to
stop databases, snapshot the filesystem, backup the snapshot, stop the
DBs, then remove the snapshot.
Julio Cesar Ody wrote:
Hi all,
has anyone got any experience with GNU/Linux snapshot file systems?
I googled around and found these:
http://www-mddsp.enel.ucalgary.ca/People/adilger/snapfs/
http://sourcefrog.net/projects/snapfs/
Didn't get any of them installed yet because I won't put my hands
(root) on any GNU/Linux OS in the next day or two.
What I want is something similar to FreeBSD's snapshot file system,
which creates a read-only mount point containing a snapshot of the
data in a section/partition/whatever, where you can copy stuff from.
Any help is appreciated. Cheers.
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
