Help with 'Failed to download the chunk' issue

Support
#1

I’m struggling with a ‘Failed to download the chunk’ problem.

This is a repo I’m not bothered about keeping the history, so I have followed the instructions on dumping the cache (I’m using the webGUI), and then running prune -exclusive -exhaustive, but it hasn’t resolved the issue and both prune and check get stuck with missing files.

I’m obviously missing something but can’t figure out what. Help!

Running prune command from /cache/localhost/all
Options: [-log prune -storage Onedrive-AppData -exclusive -exhaustive]
2025-01-22 09:39:51.701 INFO STORAGE_SET Storage set to one://Backups/AppData
2025-01-22 09:40:18.758 WARN DOWNLOAD_FOSSIL Chunk b6440e696c0171199112d6d94fecf4fd74e3c58bacedcd2aafa4b84b530b4873 is a fossil
2025-01-22 09:40:18.964 ERROR DOWNLOAD_CHUNK Failed to download the chunk b6440e696c0171199112d6d94fecf4fd74e3c58bacedcd2aafa4b84b530b4873: 404 The resource could not be found.
Failed to download the chunk b6440e696c0171199112d6d94fecf4fd74e3c58bacedcd2aafa4b84b530b4873: 404 The resource could not be found.
#2

Probably fallout of the interrupted prune. "Missing Chunks" reported following a purge - #2 by saspus

  • run check -persist
  • identify affected snapshot revisions
  • delete the affected snapshot revisions from the storage manually (snapshots/name/1,2,3…)
  • run prune -a -exhaustive to clean up orphans
#3

Thanks for the reply, but it didn’t work out. Seems like the missing files were in the fossils folder, I tried copying them back to chunks but it got stuck later on.

I’m just going to reset the repo, as the history wasn’t important and I would rather just start a fresh with a filesystem I can trust. The rest of my backups have been spot on though.

#4

How did it not work out? Did not check -persist produce list of affected revision versions?

Starting over every time something happens is not a sustainable approach going forward.

Ghost snapshots remaining in the datastore when prune is interrupted is a known issue with a known workaround. There is no data loss, just extra data that needs to be removed.

#5

When I followed the instructions, I didn’t find the missing files in chucks, but in fossils.
I’ve spent a while trying to sort this myself, and the data is a second backup, and is only short term, so starting over is fine for me.

I removed all the files and reran the backup and it looks healthy to me.
Thanks!

#6

In future, consider running check -fossils (and then with -resurrect).

Also, you can always create a fresh ID and re-use all the chunks, rather than wipe the whole thing. All too often I see people starting from scratch when it’s totally not needed, even in the most disastrous of circumstances.