Duplicates problem or errors? (Context cancelled)
Hello. I am uploading (from folders) some photos and videos for a friend and am getting lots of "file duplicated in the input" messages. But in the analysis at the top of the GUI the figures for "Duplicates in the Input" and Files uploaded do not add up to the total for the file type. When I delve into the filesystem and search for some of the named files I notice that my friend sometimes has the same file in 2, 3 or even 4 folders. Can you clarify if Immich-Go checks everywhere or does it just report as a duplicate once it has detected any duplicate in the folder structure? This might explain why my figures are so far out.
The log file also reports Err context canceled multiple times about an hour after finishing when I choose images only. When I choose videos only the figures do seem to add up and no context canceled at the end so I'm getting very confused.
What does context canceled mean and is there anything I can do?
Many thanks.
Immich-go does it best to discards duplicates in the takeout, but also it tries to not submit a photo already known by the server. Duplicates are checked with their check sum. The name and the date of capture are checked to track resized / compressed photos.
The number of assets in the input should match the number of processed assets
Input assets: DiscoveredImage + DiscoveredVideo Processed assets: Uploaded + ServerError + NotSelected + Upgraded +ServerDuplicate + ServerBetter + DiscoveredDiscarded + LocalDuplicate + MissingAssociatedMetadata
I see in the log an error that has interrupted the process in the middle:
2025-07-07 11:52:53 ERR read /media/al/TI31204800A/Users/Christine/Pictures/2014-07-16/P1000190.JPG: input/output error
Apparently this is an error occurring when reading this file. This stops all the processes (context canceled) before having processed all the discovered files.
Ah thanks. I've now deleted the corrupted photo and re-run Immich-go and this time is has completed without errors. Also this time the number of duplicates + number of uploads is pretty close to the number of scanned files.
Thanks for this great tool.
I wonder if it might be possible to trap errors like this and continue with the uploads while reporting any errors at the end of the process?
Oops just re-read the readme and see that there is an option to ignore these errors with --on-server-errors. The only thing is, the message Err context canceled is not obviously a server error, at least not to many users like me, especially as you don't see the message until after you dismiss the GUI output.
In you case, the error is coming from your local filesystem. I need to check within the source code how this kind of error is managed abusively as a server error/
Anyway, it's strange to have an IO error on only one file. Is your disk healthy?
Ah apologies. This is a corrupted photo on the source drive. When you open the picture in a photo viewer on Ubuntu you see about 70% of the photo at the top and then a black portion along the bottom. I've had a few of these with older digital cameras. Immich web uploader seems to import then without an error but then will not display them inside Immich.
Immich-go does following
- copy the file locally
- read the exif part
- upload its content on the immich server
the error message isn't precise enough to indicate which operation has triggered the error. But in your case, I guess this is exif reading that fails.
I'm planning to improve the error message