I have only one backup of my media files on mega . Using the webui to restore only gets me so far before the live log shows me that no progress is being made. Around 50% of the files have been restored. The temp folder contains 9.4/GBs. No filesets found on the dlock.zip.aes file which is chocking the restore process.
It’s possible (likely) that the power surge damaged the data on the lvm which was then backed up by Duplicati.
I tried to follow the advice to use Duplicati.CommandLine.RecoveryTool to download the backup. This fails to download any files since I am stumped by how to give —auth-passphrase the special characters in the Duplicati passphrase.
It’s possible I can restore most of the lost files, but only if I can Duplicati to get past this blockage.
I realise I am not very “ept” in using this fantastic software, but with your help I believe I will get back my lost media files.
Putting things together from two topics is probably worse.
Not clear on this reference. Is this partly referring to
which was named in one post (which doesn’t show much from single view of status)?
There is no mention over there of filesets. Filesets correspond to dlist files not dblock.
Then the latest version would have the bad data, but should still restore without issue.
Ordinarily one just uses a version before the problem, but I see the version statement:
which maybe you can clarify, as I don’t use MEGA, so don’t know the connection there.
Duplicati versions are independent of any versioning that the storage provider provides.
If you know times of the brownout, you could tell if there’s a pre-brownout backup there. Restore (if you have a database) has times. Or begin a Direct restore from backup files.
Or look at the destination files, find the dlist files, and convert from UTC to local times.
If you genuinely have only one version with only damaged data, that raises the difficulty.
Is it a special characters problem (needing shell protection) or a passphrase problem?
For example mega://folder?auth-username=username&auth-password=password.
Mega.nz also talks about these. Also note the name. It is not called auth-passphrase.
Do you have extra space on local drive? Do you have any other MEGA download tool?
I’m wondering if you should grab a local copy before you have an accident with MEGA.
Most of the issues were caused by my lacking expertise. The brownout occurred on 14 July but fortunately the Duplicati backup on mega is dated 12 July. Direct restore over the internet was always incomplete; I tried several times. Mega.nz support is mentioned on this forum. They will support their own apps, recommending not to use third-party apps.
Mega does have a cli tool: mega-cmd and using that I followed your excellent advice to download the Duplicati backup to a local disc.It has taken a while but I am happy to tell you that ALL my lost media files have been restored using the webui.
Duplicati did make the backup over the internet to mega. It was a large amount of data: hundreds of gigabytes of *.zip.aes files. From my recent experience, restoring large backups using Duplicati are best done with the help of mega-cmd.
Mega support warn that mega-cmd backup is in beta and to contact them in case of problems.
At least they seem to support it. The library Duplicati uses has recently gone nearly unsupported according to its author, and of course MEGA doesn’t want much to do with third-party software…
This is great to hear, but I’m starting to wonder if Duplicati MEGA support is feasible long term, despite your excellent finding that some of the large-restore flakiness can be worked around…
It would probably be possible to use the mega CLI tool they support similar to the way Rclone is implemented. I haven’t looked too deeply, but it seems to have basic get and put functions, which is all that is needed.
Anyway, the point is that it may be another step towards a backend based on command run.
GET, PUT, LIST, DELETE are technically the basic need, but MEGAcmd has all four of them.
I’m not seeing explicit folder creation. Maybe it’s implicit or maybe it’s easy in MEGA WEBUI.
Anyway, nice idea if someone wants to pursue. Technically rclone can already support MEGA, raising the interesting question of whether that might have worked better than our library does.
Rclone is not MEGA code. The rclone developer took on library maintenance for use in rclone. Although possibly it works better than our lib now (needs test), it still has third-party-code risks.
