Api:ErrorInterceptor~tza2l33l] Unknown error: PostgresError: there is no unique or exclusion constra
Just updated to v143 and switched to the new timeline. My device appears to be slow updating and I checked the immich logs to reveal…
350 Replies
:wave: Hey @praying for unix,
Thanks for reaching out to us. Please carefully read this message and follow the recommended actions. This will help us be more effective in our support effort and leave more time for building Immich :immich:.
References
- Container Logs:
docker compose logs
docs
- Container Status: docker ps -a
docs
- Reverse Proxy: https://immich.app/docs/administration/reverse-proxy
- Code Formatting https://support.discord.com/hc/en-us/articles/210298617-Markdown-Text-101-Chat-Formatting-Bold-Italic-Underline#h_01GY0DAKGXDEHE263BCAYEGFJA
Checklist
I have...
1. :ballot_box_with_check: verified I'm on the latest release(note that mobile app releases may take some time).
2. :ballot_box_with_check: read applicable release notes.
3. :ballot_box_with_check: reviewed the FAQs for known issues.
4. :ballot_box_with_check: reviewed Github for known issues.
5. :ballot_box_with_check: tried accessing Immich via local ip (without a custom reverse proxy).
6. :ballot_box_with_check: uploaded the relevant information (see below).
7. :ballot_box_with_check: tried an incognito window, disabled extensions, cleared mobile app cache, logged out and back in, different browsers, etc. as applicable
(an item can be marked as "complete" by reacting with the appropriate number)
Information
In order to be able to effectively help you, we need you to provide clear information to show what the problem is. The exact details needed vary per case, but here is a list of things to consider:
- Your docker-compose.yml and .env files.
- Logs from all the containers and their status (see above).
- All the troubleshooting steps you've tried so far.
- Any recent changes you've made to Immich or your system.
- Details about your system (both software/OS and hardware).
- Details about your storage (filesystems, type of disks, output of commands like fdisk -l
and df -h
).
- The version of the Immich server, mobile app, and other relevant pieces.
- Any other information that you think might be relevant.
Please paste files and logs with proper code formatting, and especially avoid blurry screenshots.
Without the right information we can't work out what the problem is. Help us help you ;)
If this ticket can be closed you can use the /close
command, and re-open it later if needed.Can you connect to postgres and run
\d asset_exif
?appears to break uploading - i signed out and back in
how do i connect again i forget haha
attach
That's not the full output
Oops I’ll try again in a few gotta put kids down
try 2 @bo0tzz

any other outputs ya'll need
So the logs output this any time im uploading
switching to old timeline shows 16 failed uploads
Error persists in v1.143.1
Hi @bo0tzz any chance folks had a min to look at these? happy to help how i can
Seems like there's some constraints etc missing from your database 🤔
Can you post the logs the server outputs when starting up?
And your .env file for completeness
Might be worth looking through these discussions as well (#20328, #20530). Can you run this on the server and post the output here?
[Discussion] Upgrade to 1.136.0 fails - missing triggers (immich-app/immich#20328)
[Discussion] PostgresError: constraint "FK_c0117fdbc50b917ef9067740c44" for table "asset_exif" does not exist (immich-app/immich#20530)
do i need to shut down server first?
try adding --prefix server
So npm run --prefix server ...
Now it's cat server/migrations.sql
attached
Have you done a server dump and restore or what is your server history?
i restored right after installing v1.143.0 bc it was having issues starting
restored from the most recent nightly backup
You are basically missing all the indexes and foreign constraints.
o
i had anotehr issue that skatbo helped me with on Github
still receiving this error in logs
So you basically need to manually fix this database drift. I still have no idea what/how it is caused. I can only guess a database dump and restore that someone doesn't create constraints or indexes.
The command you ran just detects database drift - the difference between your database and what it should do. Basically nothing will work as long as it is so far away from what it is supposed to be.
oh yeah this is me posting in this thread
So you need to go through and run the SQL in the file
And then run the command again and eventually you should get down to zero differences
how do i do this?
Do you know how to connect to the database container to run SQL?
docker exec -it immich_postgres psql --dbname=<DB_DATABASE_NAME> --username=<DB_USERNAME>
and then…
after that noThat should get you to a terminal where you can type in SQL I think
Can you do like
select * from user;
?yeah after running that command i can interact with the db
So now take the first line from the migrations.sql file and try to run it
ALTER TABLE "asset" ADD CONSTRAINT "asset_pkey" PRIMARY KEY ("id")
that?Yes

Great
Now run the command on the immich_server container again
Is the file 2 lines shorter / is that first line now gone?
thats a negative
Maybe you need to type a
;
still
Try again
ah it worked
It didn't send anything until it received a semi, so it sent the command two times
it reduced
Great
i ran the command you suggested again and see it reduced
Basically I would just work your way through the file
is there an automation for this
Lol
The automation is to not lose them in the first place
i didnt do anything int he db prior to this
You can copy/paste everything at once if you want
so now im sad
what about the --missing in target?
Don't do the // down part though
-- is a SQL comment so it's fine
? whats that
got this a few times
Oh nevermind. It used to output SQL to undo the "up" but it doesn't anymore.

these all cause issues
This one means you have duplicate assets in the table now, so it can't add the constraint.
do i ignore?
No....
Immich prevents duplicates by requiring assets have a unique checksum
You have at least two assets in the asset table with the same checksum.
this is the remaining issues
Because the constraint got dropped it didn't prevent it from being added.
This is all that's left of the output of the database diff?
yep
i ran the rest
That's good at least.
We need to delete the invalid data and then try to apply it again, for each one.
You need to do something like
select from asset where "ownerId" = is AND "checksum" = checksum;
Using the values from the errorselect from asset where "ownerId" = 54bb8e76-396e-433e-a188-6a2d1a029c2e AND "checksum" = \x8fd9a48e5387118cbdb3ee7ca5f58b641d67dc50;
?sure
does it return anything?
\
the id needs to be in single quotes at least

do
select *
from ...
How many assets/users do you have in you server?1 user, 30k ish
attached output
Yeah. The options are to go through this tedious process of of removing invalid data and re-applying the missing schema changes or make a brand new install and re-upload/process everything from scratch. The data cleanup isn't too bad once you get the hang of it.
i already deleted some of the data, so i think i have to go thru and manually fix

re-uploading is somewhat out of the mix
if you want, you can change the sql to
select "id", "createdAt", "originalFileName" from "asset" ...
You can see that the top record is probably the "original" and then you have 4 "bad" copies yesterday that were duplicates.so strange
so what do I do?
So now we want to run this
this should delete the 4 duplicate records and then you should be able to create the unique constraint
syntax error
Did you also take a recent database backup in case something goes wrong and you need to restore?
yep
i have nightlys
(14 or w/e the defualt is)
oh
it's
<>
instead of IS NOT
I updated the sql
lol
so i jsut run it back right
hold on
I wonder if you phone has uploaded like an insane amount of duplicates
it feels possible
iCloud or something glitched out a while ago
i cant even copy paste the output is so long
i dont think its worth copying, but its an endless scroll basically
this forever

this should tell you how many assets have bad duplicates
140 rows
that's not too bad at least
i cant figure out how to copy paste
but yeah its 140 rows
like this
just need to figure out if there is an easy way to delete all of them at once minus the original
you know what
maybe we should do this part last. Let's fix the other issues first and do this one last.
the other issues are easier to fix as well
but yeah, for this, we need to essentially go through one checksum at a a time and delete the rows that are the oldest
incredible
so if i start here instead
bruh
it's the same thing
oh.
pain.
chat-gpt came up with this though:
you can test it with this though
I think if this works we can probably use it for each table with problems
that damn llm
this produced 507 rows
everything was created on 9/23
that seems promising then, I'd run the delete
hold on
oh
i ran it
concerning?
it's fine no worries
now what
I was wondering if we wanted those ids to delete from asset_exif as well and the like, but it's not a big deal
try to make the UQ now
it worked
thank god
wallahi
yeah i ran that migrations.sql thing you asked before nad that line is gone now
great
pick the next one and run it and show me the error for it
so it is saying that records exist in the
smart_search
table that are not in the asset
table
So we just deleted from bad records from the asset table and it's saying those (or other old data) may have been removed, but orphaned rows still exist in that table. This should be easy to fixGo on.
sec lol
and then this again
worked
linegone
same method?
deleted 536
i ran the command and it removed the line
(alter table)
how do you come up with this?
the FK you are trying to add is linking the two tables together, so I am just putting in the right table a, table b, table a column and table b column
oh i just replace the word
nvm got it that time
yeah I changed the table once
do asset_exif and asset_face both
these both worked
this one is assetsId not assetId btw
which ones are left now?
only 3? nice
yeah
magic
hmmk
actually
do you even care about memories?
alright, the same memory_asset record exists multiple times, similar to what we did with assets
they're kinda cool
i actually use them these days
ok we can fix it
i went from disliking them to liking them after seeing my kids
haha
delete 5734
lol
it worked
?
Count 222
ok so you still have memories at least lol
222 sounds more realistic I think lol
ok, same issue with duplicate keys I guess
Same for this last one?
well
the asset_job_status doesn't have a createdAt column
it has 5 date columns for each job basically
It doesn't really matter.
maybe let's verify each one has
metadataCreatedAt
thoughhokay, gotta take a break to feed small human and i can come back
no problem!
thx for all the help sofar!
hope it fixes it
No problem. When you get back, try to run this and let me know if
metadataExtractedAt
is null
for any of the records.
no null
here
the ones without a date in the second column are null
we probably only want to keep the ones with dates tbh
Oh.
is it actually null or is it an empty string...
It’s more rows than I could copy bc idk how to use termius
Uhhh
I copy pasted
woops
sec
Assuming that still returns rows, you can run this:
returned 123
yeah, run the delete
Final boss
last one?
yeep
🫂
I think we just delete the second row, since there's no date column on the table and the data should be identical anyways
Small fluffy animal needs bathroom now
How did this all happen btw?
I have no idea, but it is quite concerning, especially if this has happened to you multiple ties now. Definitely take a database backup immediately after you get this issue resolved
delete 120
migrations sql empty
doing a backup now
i have a feeling its related to "new timeline" btw
bc thats something i did attempt to do on 9/23
i tried enabling new timeline
hmm…its missing a bunch of images that are backed up
hmm
It came up at least though?
so its really odd - my desktop (no local photos) only showing immich images from last thursday
my mobile device (where the photos are from) is showing all recent images w/ a check mark
sorry not all, but a bunch
i've got a lot of
!
images too
hmm i pressed start backup (on old timeline)Ah, I think I see the problem
the immich_server logs are clean (before they threw errors)
how many are missing thumbnails? over 100?
I'm guessing just the most recent assets are missing them. Is that correct?
Prob less than 50 (they are most recent)
yeah, there was a slight issue with sql to clean up duplciate ids, which we ran on memory_asset and asset_exif that deleted both records, instead of just one (I don't even know how to fix this problem when the primary key itself is duplicated)
Since it's just asset_exif (which I'm pretty sure is affecting the query for the thumbnail), we can regenerate metadata for those
But that doesn’t solve why it’s not showing any assets after last Thursday tho? On the web client
Just select the most recent picture on the web, scroll down until you are past all the broken thumbnails and shift select another one, then use the bulk actions to run "Refresh Metadata"

I just run this in admin jobs
The most recent web pic doesn’t include the no thumbnail on mobile
oh I see
let me check something
dang, yeah. the timeline on the web won't show anything without exif records and we accidentally deleted like 100 records or so
if you run metadata extraction on everything that will fix itself
Kk
hmm

thats a bigger number
100k assets?
I thought you only had like 30k lol
look man maybe i lied

I hope it's not 30k duplicated
the disk usage is close to being correct
which is whats confusing to me
i assumed if it duplicated, it'd be double the space
its off by maybe…30-50gb
max
we deleted about 500 images that were uploaded again
those records in the database were removed, but the original is still going to be there on disk, there's not a super easy way to fix that atm
we're hoping to build out some features in the future which will make it easy to detect orphan files and clean them up though
its making my cpu happy

lol
I have no idea if it starts at the beginning or the end
its ok i might need to sleep on it while the job runs
should i try to enable new timeline
or should i wait
I'd leave the mobile app alone for a bit
ok i'll let this server cook
Let the jobs finish, verify everything is fixed, take a dump, and then turn it on
Is your database server on a drive on the host or a network share?
host
SSD?
nvme yep
This has happened twice already?
uh the other time was way easier to fix
in that i did whatever is in that github issues page
and it was resolved
did you run the script to generate the diff or just fix the error to get it to start?
uh i just fixed the error i think
if you follow the chronology in that thread, i just did literally whatever was posted first
and then whenever it started working i moved on with life
and here we are
yeah makes sense
so i probably put a bandaid on this bigger problem
so those foreign keys and indexes might have always been missing ever after you fixed the one that was causing a problem originally
that'd be less concerning tbh
which thread was it again?
I want to see if the same constraint came up again or not
[Discussion] PostgresError: constraint "FK_c0117fdbc50b917ef9067740c44" for table "asset_exif" does not exist (immich-app/immich#20530)
FK_c0117fdbc50b917ef9067740c44
was missing?
that was renamed to asset_exif_assetId_fkey
which you had a problem with today[Discussion] PostgresError: constraint "FK_c0117fdbc50b917ef9067740c44" for table "asset_exif" does not exist (immich-app/immich#20530)
this has the schema
(the original post is not mine
i just piled in bc similar)
ok its not fast, gonna need 2 sleep and i'll report back

[Discussion] PostgresError: constraint "FK_c0117fdbc50b917ef9067740c44" for table "asset_exif" does not exist (immich-app/immich#20530)
yeah, looks like you fixed these 3
and there was just many more
?
That particular migration made some database changes and the changes failed without the constraints/indexes being there
Basically we ran SQL that uncovered specific problems
I see - this is very fun, i hope it helps someone else.
But those problems were different than the ones today, so it still could have been a single event that happened sometime before that
Anyways, I would probably recommend running that diff command periodically to make sure stuff is still there
Let me know tomorrow if you are still having issues! Have a good night!
from overnight logs
Pretty much all new assets (not in immich) are showing no thumbnail in the mobile app now
Mobile app is now being excruiciatingly slow (even after pressing start backup, nothing is happening)
took out lines w/ albums
so the mobile app is still showing broken thumbnails for many assets, but at least web client is now showing up to Saturday.
Missing Sunday to Wednesday now…
Again, the strnage thing is mobile app shows the asset is backed up (check mark) but the assets dont appear in web
I'd try the beta timeline now
Checked the logs bc the thumbnails were still broken
#0 _futurize (dart:ui/painting.dart:7991)
#1 ImageDescriptor.encoded (dart:ui/painting.dart:7785)
#2 instantiateImageCodecWithSize (dart:ui/painting.dart:2558)
#3 PaintingBinding.instantiateImageCodecWithSize (package:flutter/src/painting/binding.dart:147)
#4 RemoteImageRequest._decodeBuffer (package:immich_mobile/infrastructure/loaders/remote_image_request.dart:164)
#5 RemoteImageRequest.load (package:immich_mobile/infrastructure/loaders/remote_image_request.dart:32)
<asynchronous suspension>
#6 CancellableImageProviderMixin.loadRequest (package:immich_mobile/presentation/widgets/images/image_provider.dart:59)
<asynchronous suspension>
#7 ImageStreamCompleter.setImage (package:flutter/src/painting/image_stream.dart:724)
<asynchronous suspension>
oh okay uploading is now working in new timeline
the logs are relatively clean
It’s just stuck here

can you click view details?
No uploads in progress
is your mobile app on 1.143.1?
It might be backing up still and the ui is just not updating, or, it might be frozen and force-closing and reopening the app might fix it.
just got update
same behavior
assets in view details are backed up tho
Ok noticing a pattern on a few. They’re videos that are short but the app is showing them as .HEIC
Also the web client seems to be not showing video backups between Sunday and Wednesday
videos might be done last
they also might be live photos
just wait until everything is backed up and then see if there is still an issue between mobile and web
but the app is showing check marks on stuff, doesnt that mean backed up?
I have way fewer broken thumbnails now, but the web client just isn't showing some stuff
OK weird behavior…
I saved 3 videos from a different device (my wife took them and sent them to me) - those uploaded fine
I took a new vidoe now - it's stuck in the "Remainder" count - now at 136 instead of 135.
And my video is definitely NOT backed up, but the one's i saved from my wife are
Her videos have a check mark, my video has a crossed out cloud
logs attached
for completeness

spotted in my logs just now
OK - narrowing in on its the New Timeline…
I just switched back (disabled New Timeline) and the new items uploaded (no idea about those other 135)
Old Timeline still has a bunch of broken Thumbnails (that were fixed in New Timeline)
hope this helps w/e going on w/ new timeline
i just want uploads to work (lol) - so its okay if i dont have New Timeline for now
No database drift by switching between (just wanted to confirm)
Btw on old timeline shows 30k duplicate assets (idk why)

Adding to the weird behavior - I now have many recents folders with various dates

So some of the problems might have been caused by the broken constraints, which let you upload duplicate images and we've deleted some of those
It might be worth logging out and back in
already tried, took those logs/screenshots after wards
54bb8e76-396e-433e-a188-6a2d1a029c2e/54/92/54920d1d-f374-4165-aecb-63eb4071af7c.MOV
It might be worth navigating to
/photos/54920d1d-f374-4165-aecb-63eb4071af7c
on the web
You seem to have 2 different issues:
1. Files not uploading from phone
2. Broken thumbnails corresponding to files uploaded to the server that aren't actually there
Files upload from phone on Old Timeline, but do not upload from phone on New Timeline
Also, Thumbnails are broken on New Timeline, but functional on Old Timeline
This just may be open issues with the new timeline then
Is anything still broken on the web?
web client still is not showing any images from Sunday to Wednesday, but is showing all new images/videos
Sounds like the old timeline is also broken in a different way then as it's not backing up those pictures
Are those the pictures the new timeline refuses to upload?
it doesnt appear to be related to that. New Timeline shows those 135 assets (mix of recent, old and very old) that won't upload. Then any new assets (after the 135) won't upload at all.
New Timeline makes all the broken thumbnails appear as normal
So they are failing to backup on both timelines?
No Old Timeline succeeds at backing up the newest assets (the 135 are still a mystery to me, they're backed up - they've been backed up) but they show in new timeline as in queue
It sounds like they are not actually backed up, especially if you cannot find them on the web
No no - the assets not showing on web are specifically just assets uploaded on the date of this past sunday to wednesday
the mysterious 135 assets do show up in web
Ok I don't really care about the old timeline at all
There are 135 assets that are not showing up on the web
And a different set of assets not having thumbnails on the timeline?
the 135 assets (that new timeline has in queue) don't seem to have an issue; i'm going to switch one more time and see if i can triple check this
It would be good to figure out why they are not progressing
what steps can i take to figure out why
i tried a bunch of intuitive things like signing out lol
Just the logs honestly
ok so i checked multiple images - they common issue with many of the remainder items is that they're 2ish second videos uploaded as
*.MOV
BUT the Remainder logs shows *.HEIC
the really common piece is video length 0:02sounds like live photos
do you have an iphone?
they are actual videos
ya its an iphone
but i know for sure, these are short videos
do you take live photos?
i take intentionally short videos (and live photos)
and these are definitely not live photos?
yep
they're definitely actual videos (the files in the server are MOV) but the "Remainder" log in mobile app for some reason is showing HEIC
can you send a screenshot?




all are 0:02

Looks good.
And there is nothing recent in the logs that would indicate why videos are not being backed up?
2025-09-25 13:53:43.405863 | warning | StorageRepository | Error getting motion file for asset 7BC5C3A5-CC74-4771-BBC5-2234E79DED2A/L0/001, name: IMG_1449.JPG, created on: 2024-12-21 14:13:13.000 | PlatformException(PHPhotosErrorDomain (3169), The operation couldn’t be completed. (PHPhotosErrorDomain error 3169.), No failure reason provided, null)
3169 is network error, presumably, the asset is on iCloud and the app is struggling to download it. From the error, this also looks like a live photo. Can you find the asset on the photos app using the metadata from the above log and check if you can share it from there?is there any way to search immich for file name like this
Oh you've to search the file on your iOS Photos app, not within Immich
You can use the date to narrow it down
i cannot even find this file name ot save my life
Can you locate the asset inside Immich using the "View in timeline" button from the Remainder list?
oh i see now in the remainder list, it's in a "Shared Icloud" album
You can enable Advanced Troubleshooting in the app settings, and once you have the asset, long press it, tap the "Advanced Info" button in the bottom sheet that opens. Should give you more metadata about the asset and the album that it is a part of
aka one i dont own directly, but i have access to
From the Photos app, Can you see if you can share / access the asset?
Yeah i can see it
Can you play the live video as well? Try sharing it from the Photos app and see if the "Preparing" dialog ever finishes
It plays!
If it plays now, Can you try going into the Immich app, and toggle the backup button and see if the upload queues gets cleared
still stuck
Would you mind sharing the app logs again? Just want to confirm if it is the same error
Hmm, odd. The same errors. Also, the reason that this isn't showing up on the old timeline could be due to the fact that the old timeline logs these errors and moves on without processing them
lmk i had to step away for a while today in zoomland
im considering going thru each 135 asset and deleting them lmao
its working to decrease the remainer
by eliminating the asset.
the strange part is many of these assets have a duplicate that is uploaded
OK so i removed "Shared iCloud Albums" and upload started workign again
im gonna try adding them back
OK i found the issue
the Shared iCloud assets are already uploaded or look to be but if I open the asset (and play video) i can "click Upload" and force the upload
and the asset moves on
its almost like you can just force it to try again and it works
Okay I just removed the iCloud shared album and uploads are fine from local recents
Nice!
id like shared icloud to work; only one album is having issues
would be nice to force continue
I think we actually deleted memories
I probably had thousands bc I have an image or video or more everyday for the last 13 years
We can regenerate them easily enough if you want. In any case we only generate memories 3 days in advance, so they will fix themselves probably day after tomorrow
https://github.com/immich-app/immich/issues/16484#issuecomment-2692718993
Just replace
memories
with memory
and it should workOnce the upload goes through all the assets from the Recents album, you can try adding the shared album as well
I tried it’s such on 8 assets now that I can’t force upload
No log errors either
Just literally won’t upload
Are those the ones you’ve received from a different iCloud users or are they your own assets but on a shared album?
They're a shared album owned by someone else that i joined (that i contribute to) but the specific assets are owned by someone else and uploaded by someone else
The trickier thing might also be some items I took the picture (thus are in my recents) but sent to someone else who then uploaded to said album
I see, Thanks for the info! Can you download the problematic assets from the shared album to your device and try uploading them from there?
1) I take picture
2) I send to waifu
3) Waifu uploads to shared album
4) Immich confused
Or through the web, as long as it is the same asset and the checksums match, the mobile app will skip them even if they are uploaded through other means
ok i can…attempt…its really weird bc i can open the asset on Immich mobile but for these specific ones when i press upload
literally nothing happens
All of them seems to be a live photo. When you open them, you're viewing the photo part of it. The way the live photo upload works is to first send the video part of the asset followed by the photo part. And since we have trouble accessing the video part, it sits there without doing anything
YO
I airdropped from imich mobile app
and uploaded via web
and we have moved on.
:ppParty:
its so fucking hard to find the photo tho - the little icons are like a scavenger hunt
wallahi my brothers we have fixed it
a support team of devs hates to see it

No problem lol! Glad we sorted things out 😃
wait, so what was the issue then?
cannot access motion parts of a live photo from the iCloud shared album?
we need to fix this though, that should not block/prevent other uploads from progressing. do we at least skip them now? maybe we should at least add something to the logs when this happens so the user is aware of what is going on
We do have the logs. But we might not be skipping them
Yep, the actual error is thrown from iOS side, so there's not really much we can do other than asking the user to upload it through the web
We'll should definitely address the blocking issue though!
yeah, the blocking issue, and also feedback as what's happening or why it's skipped
Oh I see I completely missed this in the excitement of fixing it
tbf i'd accept a toggle that says backing up live photos may result in only still images in edge cases
idc that much, i do love a good live photo
The airdropped one did retain the motion part of the image as well, right?
Yep cuz they’re heic
I’m pretty sure
I deleted them and they were already backed up so idk
Don’t make me look again
I have the memory of a goldfish
lol alright
It doesn’t explain the other weird MOV ones
Where remainder claimed the file was HEIC
But the file was indeed a 0:02 video
That was a bizarre case
I swear to you I never change Live Photo to video
You still have those in the remainder? Or do they come back if you add the shared album to the backup list?
I removed them from device
Ah, I see :monkaHmm:
They were local files
In recents
I have a feeling however you’re looking for heic/videos it’s looking for a specified length and I just happen to have that dog in me to record the exact length of a Live Photo
Were you able to play the videos through the app though?
Yup they play
Remainder just kept saying hey there’s an heic file I can’t upload
And I was like
It’s MOV
and here we are
This error specifically
I just saw the resolution
And now I’m not even confident they were videos
Wtf
I hate it here
Anyway my solution for local files was just to delete the file
So I can’t even tell you now
But what you said wrt Live Photo motion makes sense
Feel free to ping me if you have anything weird with the mobile app
It just randomly started backing up 419 assets that were already backed up

Doesn’t look like it’s getting stuck
OhNOOOOOO
it’s telling me the same assets as last time haven’t uploaded again
Are these the same assets? Can you close and reopen the app once and share a screenshot of the backup page
Yeah they look like the assets I “fixed” before

Okay another complexity. An asset exists in two different shared albums
And it seems that’s tripping it up
1) I take video
2) send to Waifu
3) Waifu uploads to Album A
4) Waifu also uploads to album B
5) I attempt to backup both albums
6) Immich sad
I can’t tell why (after doing nothing) it’s asking these be uploaded again?
Just to throw it out - some assets are JPGs (not Live Photos) that are getting suck
Inside Recents (so local file) but probably sitting in iCloud bc of “optimized” storage
Wait, Can you update your app to the latest version?
143.1?
The latest one has a new UI that shows the number of assets that are being processed and those that have been processed
Yep
i am on
oh yeah i saw that - they are on the right side of the UI already
instead of "preparing"
I think it is displayed only when hashing is in progress, so if it is not being rendered, then we can assume that everything is hashed
Let's try this then, do you have the advanced troubleshooting enabled?
i can
its enabled
Before checking this, does the timeline display two different thumbnails for the same asset? Can you find any duplicates there?
yeah i can
all the ones that are "stuck" in remainder are duplicates
in that the file is already uploaded
or a version i guess
Then can you share the screenshot of the advanced info page of the one that is being tried to be uploaded and the one that is already on the server?
oh no
some voodoo just happened
and everything just uploaded
🪄
it went from 31 to 0
but you should still be having duplicates on the timeline, right?
no once i upload or delete the problematic asset
it stops being a duplicate
It doesn't make sense lol, for it to stop being a duplicate, it has to have the same checksum, and if it has the same checksum as the one on the server, it should not be in the upload list at all, since the server already has it
you're telling me boss
If at all this happens the next time, Grab a screenshot of the Remainders page, as well as the advanced info page of both the assets
kk
It might help us troubleshoot what's happening with your device. For completeness sake, Your server is also on 143.1, right?
yup


Same asset
I ain’t time traveling
7BC5C3A5-CC74-4771-BBC5-2234E79DED2A_L0_001_1758825782.100549_o_IMG_4289
full file name after airdrop
okay its definitely duplicated in my library now
tremendous i love it.
OH I remember why it might have done the ask for the assets that got blocked.
I used deduplicate on the web app
So I bet I removed the problem assets
Yeah it was definitely this deduping
I just checked duplicate tool - and its showing 300 duplicates, i bet if i delete these - i'll have 31 blocked assets agai
Took me a whileto realize - you should somehow account for the checksum of a deduped file then tell mobile app to pls not attempt reupload if its been dedupedOh yeah, we still have this behaviour where the app re-uploads stuffs after being deleted from the web. We’d have better handling in the future, but the way to workaround this is to also remove the assets from the device after deleting from the web app
But impossible with shared iCloud bc I don’t own the asset
That is true. Probably the best way to go about this is to let the app upload all the assets from the shared album, then remove it from the backups so they won't be retried. But then when you want to upload new assets from the album, you can go to the Library tab -> On this device -> Shared album, select the new assets manually and do a manual upload of them
My workaround would work on Android or your own assets, but sadly, might not be a solution for assets from the shared albums
Wow guys new timeline is very good when it works for me now