DigitalOcean/image migration

from HTYP, the free directory anyone can edit if they can prove to me that they're not a spambot
Jump to navigation Jump to search

In order to maintain backups of disk images without having to pay DigitalOcean for them monthly, it would be useful to be able to download them and then later upload any that need to be used again.

At the present time, downloading does not seem to be possible; the only instructions I have found so far not only don't work, but couldn't even work in theory as I understand it -- in that they use sshfs to mount the volume's filesystem locally, then try to use dd (a volume-copying tool) to do the copying.

Tentatively, uploading/restore is fully supported, although this is still limited to images that are based on a Linux-like filesystem.



I got an error with the command they suggested, and modified it to this:

root@statler:/mnt# sudo sshfs -o allow_other root@ /mnt/droplet

(If doing this on a Macintosh, see /backup/MacOS.)

...but then this didn't work:

dd if=/mnt/droplet
dd: error reading '/mnt/droplet': Is a directory

Tentatively, this seems to be the way to go, for an image that's running as a virtual machine.

1. Find the device name for the root filesystem:

Filesystem     1K-blocks     Used Available Use% Mounted on
udev             1530884        0   1530884   0% /dev
tmpfs             308032    31616    276416  11% /run
/dev/vda1       50758760 12572412  38169964  25% /
tmpfs            1540156        4   1540152   1% /dev/shm
tmpfs               5120        0      5120   0% /run/lock
tmpfs            1540156        0   1540156   0% /sys/fs/cgroup
/dev/sda        51475068 30045492  18791752  62% /mnt/cloud4-01
tmpfs             308032        0    308032   0% /run/user/0

This tells me that my main partition is /dev/vda1 and that I need 50,758,760 kb (50 GB) of space to store the image. Presumably, then, the entire (bootable) drive is at /dev/vda>.

2. cd to a local drive with enough room.

3. Download the image:

Adapting these instructions:

  • Model: ssh user@remote "dd if=/dev/sda | gzip -1 -" | dd of=image.gz
  • Live example:
ssh "dd if=/dev/vda | gzip -1 -" | dd

This takes a long time, so you may want to check up on the size of the output file (in this case, to see how it's progressing.

Also, copying an image while the disk is in use (which seems to be unavoidable with Digital Ocean) may cause inconsistencies in the resulting image, so running fsck on it afterwards, and then manually spot-checking the contents for sanity, is probably a good idea.

(Currently waiting for my own image to download; will then be able to see if this worked.)

not checked



Possibly the "custom images" feature is the way to go?