Re: [Xen-users] Just some (probably useless) information

[Grant McWilliams <grantmasterflash@xxxxxxxxx>]

On Sun, Mar 21, 2010 at 12:57 PM, Brandon Lamb <brandonlamb@xxxxxxxxx> wrote:

>> dd disk -> disk = 36 minutes
>>
> XCP also supports vhd snapshots.
>
> -- Pasi

I think there may have been some confusion. The lvm partition was 20
gigabytes, the USED space was 2 gigabytes. So for dd it had to copy
all 20 gigabytes to another 20 gigabyte disk/file whereas using the
ntfsclone only copied the USED 2 gigabytes.

I dont think snapshots would really help in this case because when a
backup of the disk is made, it may not be backed up again for a week,
or a month, or a year which I dont think works with the snapshot
model? I am also not doing a live snapshot/copy, the guest vm will be
shutdown then the disk or file copied to another disk/file.

I'll try the blocksize tip and see what that gets me.

Also try ddrescue which is as fast as dd with ALL of the optimizations turned on. With my experience and the speed in which you say your RAID is functioning you should be able to ddrescue your 20 GB LV in about 4 minutes or less.

One thought though would be this, why is your partition 20GB when you're only using 2GB? Wouldn't it be easier to have a small OS partition and use more traditional network storage (NFS/Samba) with more traditional backup techniques? If the OS drive was say 5 GB you could image the LV bit by bit in about one minute and then use rsync (or other) to backup the changes.

Grant McWilliams

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users