

Maybe it would have found something and I was just being impatient, but I don't know how it would know which sparsebundle to mount and examine as a time machine destination. It didn't find anything on the network (or maybe I didn't wait long enough), so I had to manually map the AFP share and mount the disk image from the Terminal, which was no big deal. I installed the new SSD, used a Lion Recovery USB stick to get a recovery partition onto the drive, and attempted to TM restore after booting from that. I did a restore over netatalk (to a solaris/ZFS server) without issue when I swapped out the HD for a SSD in my mini.ĭid the restore stage where it was "Searching For Time Machine Backups" find a restore source on the network? I figured I'd test my setup by upgrading the SSD in my 2009 MBP, as the OEM Samsung SSD was getting extremely slow (been in use a looong time). I've had bad luck with time machine in general (every 3-4 months it decides my time machine backup is "corrupt" and wipes history/restarts) - but this was the same over netatalk, iSCSI, and local firewire disk. You should be fine, but of course it's always recommended to have more than one backup, particularly if you're using a configuration that isn't officially supported by the vendor.

The issue PDonovan had with a backup being suddenly seen as invalid can happen even with a Time Capsule or shared OS X drive, unfortunately sometimes the backups do get corrupted in some way, or a bug in backupd requires you to start a new backup.
#Netatalk time machine backup failed upgrade
A major OS upgrade is precisely when you want your backups to be most available and up to date! That doesn't happen so frequently that I think anyone on ARS would be bothered by it, but if you're setting up a NAS for a non-geek, it can be a big consideration. Time Machine almost always requires whatever new AFP features get enabled, so until Netatalk is updated, it won't work seamlessly.
#Netatalk time machine backup failed update
It usually works great until Apple has an OS update that changes AFP behavior in any way, then people complain to Apple that Lion "broke" their NAS. Is it going to turn out to be a horrible, buggy mess, or is it generally fairly well behaved? I figured I'd poll the crowd here for experiences/tips/advice/warnings/derision about using this approach for Time Machine backups. I didn't have fool around with Time Machine, unsupported volume hacks, sparsebundle copies, etc - it all went quite smoothly. Just finished setting up a Ubuntu 12.04 VM with netatalk, and I've got Time Machine on 10.8.1 talking to it and performing an initial backup with no issues so far (much more testing to be undertaken, of course).
