Repeatable Chunk check failure with google drive

I notice that my Check job had failed in duplicacy web and reported missing chunks in the latest 2 snapshots numbers 10 and 11.

As suggested in the “fix missing chunks” page I deleted the folders for those two snapshots from my google drive under the snapshots//10 and 11 and ran the backup job again. The job seems to run ok picking up 10 as the next snapshot to use.

Running backup command from C:\Users\Phil/.duplicacy-web/repositories/localhost/4 to back up F:/Data/Phil
Options: [-log backup -storage -DC -stats]

2019-08-27 07:59:21.345 INFO REPOSITORY_SET Repository set to F:/Data/Phil
2019-08-27 07:59:21.345 INFO STORAGE_SET Storage set to gcd://Backups/<hostnamereplaced>-DC
2019-08-27 07:59:35.460 INFO BACKUP_START Last backup at revision 9 found
2019-08-27 07:59:35.461 INFO BACKUP_INDEXING Indexing F:\Data\Phil
2019-08-27 07:59:35.461 INFO SNAPSHOT_FILTER Parsing filter file \\?\C:\Users\Phil\.duplicacy-web\repositories\localhost\4\.duplicacy\filters
2019-08-27 07:59:35.461 INFO SNAPSHOT_FILTER Loaded 1 include/exclude pattern(s)
2019-08-27 07:59:44.331 INFO UPLOAD_FILE Uploaded Tech/Radio/Baofeng_UV-82_20171002-xxxx.img (6653)
2019-08-27 07:59:44.331 INFO UPLOAD_FILE Uploaded Tech/Radio/Baofeng_UV-82_20171002.img (6657)
2019-08-27 07:59:44.331 INFO UPLOAD_FILE Uploaded Tech/Radio/UV-82-phil.img (6653)
2019-08-27 07:59:44.382 INFO BACKUP_END Backup for F:\Data\Phil at revision 10 completed
2019-08-27 07:59:44.382 INFO BACKUP_STATS Files: 29972 total, 213,577M bytes; 3 new, 19K bytes
2019-08-27 07:59:44.382 INFO BACKUP_STATS File chunks: 42541 total, 213,577M bytes; 0 new, 0 bytes, 0 bytes uploaded
2019-08-27 07:59:44.382 INFO BACKUP_STATS Metadata chunks: 4 total, 12,272K bytes; 0 new, 0 bytes, 0 bytes uploaded
2019-08-27 07:59:44.382 INFO BACKUP_STATS All chunks: 42545 total, 213,589M bytes; 0 new, 0 bytes, 0 bytes uploaded
2019-08-27 07:59:44.382 INFO BACKUP_STATS Total running time: 00:00:20

However to me that reads like it found 3 new files but didn’t upload anything.

A new check job then fails with the same issue.

Running check command from C:\Users\Phil/.duplicacy-web/repositories/localhost/all
Options: [-log check -storage -DC -a -tabular]

2019-08-27 08:03:16.164 INFO STORAGE_SET Storage set to gcd://Backups/<hostnamereplaced>-DC
2019-08-27 08:03:18.682 INFO SNAPSHOT_CHECK Listing all chunks
2019-08-27 08:12:00.310 INFO SNAPSHOT_CHECK 13 snapshots and 135 revisions
2019-08-27 08:12:00.311 INFO SNAPSHOT_CHECK Total chunk size is 603,568M in 127006 chunks
...
2019-08-27 08:12:03.630 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 1 exist
2019-08-27 08:12:03.737 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 2 exist
2019-08-27 08:12:03.846 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 3 exist
2019-08-27 08:12:03.947 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 4 exist
2019-08-27 08:12:04.063 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 6 exist
2019-08-27 08:12:04.167 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 7 exist
2019-08-27 08:12:04.266 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 8 exist
2019-08-27 08:12:04.368 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_Phil at revision 9 exist
2019-08-27 08:12:04.785 WARN SNAPSHOT_VALIDATE Chunk 83ef51bacd3e1404a996422c3d008c19084a32fb6cf051a5be0f8eb591df2f51 referenced by snapshot AWDGamer_Phil at revision 10 does not exist
2019-08-27 08:12:04.797 WARN SNAPSHOT_CHECK Some chunks referenced by snapshot AWDGamer_Phil at revision 10 are missing
2019-08-27 08:12:04.800 INFO SNAPSHOT_CHECK All chunks referenced by snapshot AWDGamer_PhilVideos at revision 1 exist
...
2019-08-27 08:12:04.938 ERROR SNAPSHOT_CHECK Some chunks referenced by some snapshots do not exist in the storage

Any idea what is going wrong? The other backup jobs all seem to be working ok.

Does this chunk exist in the storage? The path should be 83/ef51bacd3e1404a996422c3d008c19084a32fb6cf051a5be0f8eb591df2f51.

Er, yes and no. In one of the 83 folders it does and in the other one it doesn’t. The real question is how did I end up with two folders both apparently with the same name and why does Google Drive even allow such a thing.

Oh I’ve heard about this. Sounds like it’s unique to Google Drive and happens for some unknown reason.

Enough so that even Rclone has its own command to deal with the situation:

https://rclone.org/commands/rclone_dedupe/

Perhaps you can use this to fix the structure of the backup storage?

2 Likes

Thanks that tool looks useful for the future.

As I only seem to have the one duplicated directory name I’ve moved all the files from one (the one with the missing chunk in it) to the other and deleted the empty one.

The check now completes so we’ll see whether the next backup screws it up again.

1 Like

I believe this was caused by a race condition when creating the same directory 83 from different computers at the same time. Google Drive does allow directories or files with the same name to co-exist, as long as each directory or file has a unique id.

However, I don’t know why the check failed for revision 10 but not for revision 9 – obviously they have the same set of chunks.