OSError: image file is truncated (paperless)

[mystereit]

hi from italy

I was sent here from the paperless github since i get an error importing some pdfs. i never had any troubles but since a couple of weeks i get the error above especially with larger files.
my scanner is a brother ads2700 but i also got the error with an epson mfc scanner. i checked the generated pdf file but it seems valid and legit.

i added this to papeless.conf and it worked for one file (smaller) but i still get the error with the 72 pages file
Environment="PAPERLESS_OCR_USER_ARGS={"optimize": 0}"
Environment="PILLOW_LOAD_TRUNCATED_IMAGES=1"

here are the logs for the "corrupted" file

ocr.txt

i hope anyone can help me or point me in the right direction

thank you very much
moritz

p.s. ofc both lxc container and paperless are latest version

[jbarlow83]

This sounds like the same Ghostscript bug as lots of people are running into:
https://bugs.ghostscript.com/show_bug.cgi?id=708956
The fix is recent, but I don't know if it's been released by them yet.

Requires changing to --output-type pdf (to avoid Ghostscript) or changing the version of Ghostscript to an older one than 10.6 but newer than 10.3 (to avoid catastrophic bugs in 10.0 through 10.2) or the end of the 9.x series. I'm not entirely how to achieve either in paperless.

[bingobob]

I have (I think) been hitting this as well. Would very much like a fix as paperless is just so goooood!

Brother scanner, latest firmware on it.

Latest paperless-ngx installed via proxmox helper scripts.

I have posted issue in another paperless comment on GitHub.

[t0rb3n]

I compiled and installed GhostScript manually since the fix isn't in any release yet and I can tell that for me the fix works.
Look at my comment in paperless repo

[mystereit]

apparently it is a problem with a specific version of ghostscript (10.6)... so i thought of simply sitting it out until a new version gets into the paperless release. what i wonder is that not more encounter this problem... and it is not only our brother scnners... i have encountered the error also with pdf generated files Am Di., 6. Jan. 2026 um 08:46 Uhr schrieb bingobob ***@***.***

…

: I have (I think) been hitting this as well. Would very much like a fix as paperless is just so goooood! Brother scanner, latest firmware on it. Latest paperless-ngx installed via proxmox helper scripts. I have posted issue in another paperless comment on GitHub. — Reply to this email directly, view it on GitHub <#1609 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BDXFXT4EMZI7TZEMSVXHH634FNR6DAVCNFSM6AAAAACPUKKQXKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTKNBSGE3TIOI> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[bingobob]

makes sense,,, yes i am currently running the "fix" - proxmox helper script LXC... would like to remove the fix from the config file when i can as it doesn't feel very good, at least it scans though and ingests... my wife scanned something yesterday and it worked....

root@paperless-ngx:~# tail /opt/paperless/paperless.conf 

# Binaries

#PAPERLESS_CONVERT_BINARY=/usr/bin/convert
#PAPERLESS_GS_BINARY=/usr/bin/gs


# fix 
PAPERLESS_OCR_USER_ARGS={"optimize": 0}
PILLOW_LOAD_TRUNCATED_IMAGES=1
root@paperless-ngx:~# 
root@paperless-ngx:~# 
root@paperless-ngx:~# ghostscript --version
10.00.0
root@paperless-ngx:~# 
root@paperless-ngx:~# 
root@paperless-ngx:~# 

[mystereit]

that fix didn't work for me :( Am Mi., 7. Jan. 2026 um 09:13 Uhr schrieb bingobob ***@***.***

…

: makes sense,,, yes i am currently running the "fix" - proxmox helper script LXC... would like to remove the fix from the config file when i can as it doesn't feel very good, at least it scans though and ingests... my wife scanned something yesterday and it worked.... ***@***.***:~# tail /opt/paperless/paperless.conf Binaries #PAPERLESS_CONVERT_BINARY=/usr/bin/convert #PAPERLESS_GS_BINARY=/usr/bin/gs fix PAPERLESS_OCR_USER_ARGS={"optimize": 0} PILLOW_LOAD_TRUNCATED_IMAGES=1 ***@***.***:# ***@***.***:# ***@***.***:# ghostscript --version 10.00.0 ***@***.***:# ***@***.***:# ***@***.***:# — Reply to this email directly, view it on GitHub <#1609 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BDXFXT2TCMEJ7AP3XLQIERL4FS5Y5AVCNFSM6AAAAACPUKKQXKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTKNBTGE3TQNI> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[jbarlow83]

ocrmypdf v17 makes it possible to avoid Ghostscript entirely for those still struggling with it.