Troubleshooting

HTTP 500 Error During Import

Typically due to Paperless updating classification models and not rebuilding the classification_model.pickle object:

logs

[ERROR] [paperless.classifier] Unknown error while loading document classification model

# Remove the classification model and restart all services.
rm paperless_ngx_cfg_data_dir/classification_model.pickle
systemctl restart paperless_*

PDF's fail to ingest with older ICC profiles.

Some PDF's with older ICC profiles may fail to be ingested. Though rare, these can be manually pre-processed to fix the ICC profiles:

gs -o output.pdf -sDEVICE=pdfwrite -dPDFSETTINGS=/prepress input.pdf

PDF failed import from consumption directory.

Paperless does not clean cache aggressively and TMPFS is typically cleared only on boot. In cases where there are mass processing of documents it is better to use disk. Some paperless subprocesses will always use /tmp.

Check logs for specific errors. Increase backing space if necessary and restart the machine or the consumption service:

systemctl restart paperless_consumer.service

Tasks may also be restarted in the django admin interface

Settings ➔ Open Django Admin ➔ Paperless tasks

# The service may also be stopped and and **/tmp** cleared
systemctl stop paperless_*
rm -rf /tmp/paperless*
rm -rf {PAPERLESS_CONVERT_TMPDIR}
systemctl start paperless_*