A list of known problems which won’t be fixed.
Case sensitive search for unicode text with SQLite¶
Warning about missing Tesseract in papermerge_app docker container¶
If you use docker-compose file provided by papermerge repository, you will notice following error in papermerge_app docker container:
WARNINGS: |papermerge_app|: Papermerge can't find tesseract. Without it, OCR of the documents is impossible. |papermerge_app|: HINT: Either it's not in your PATH or it's not installed.
This is very misleading, but it is correct and has no any impact on papermerge operation.
Above mentioned error is safe to ignore only (and only) when it runs in
main app container i.e. in
papermerge_app). Exactly same warning inside
papermerge_worker will be an error an application won’t function as
the OCR is performed by the worker (papermerge_worker docker container) and thus is installed only on the worker. However, for simplicity sake both worker and app (papermerge_app and papermerge_worker) use same code. From application code point of view they are 95% identical and use exactly same methods for reading configurations.
- for worker - tesseract is mandatory
- for main app - tesseract is optional
as long as that “missing tesseract” comes from papermerge_app docker container - you can just ignore it.
Current setup (even though little confusing) is very practical. The practicality of it is that worker and app many times run on same machine, from same folder and use same configuration file - e.g. in development mode - in such cases, issuing that warning reminds developers to install tesseract.
Slow increase in CPU usage over time¶
In case you use default broker and message queue (which is filesystem based) over time your CPU usage will increase. Filesystem based message queue is not suitable for production. Use redis to avoid this issue. Learn in broker configuration part how to configure redis.