This is a high-level overview of the currently unresolved issues. PhotoPrism follows a zero-bug policy, which means that we do our best to fix any issues we learn about. However, sometimes this is not immediately possible, for example due to dependencies on third-party libraries, other apps, or because it is a specific use case that we do not support at this time.
In order to improve readability and reduce maintenance, minor issues and those we plan to fix in the short term are not listed here. You can browse all reported bugs, ideas, and planned enhancements on GitHub Issues.
Installation in a sub-directory on a shared domain is generally possible if you run your instance behind a reverse proxy. At the moment, this method is experimental. We do not recommend it because technical expertise is required and a number of specific issues still need to be addressed:
Nested Import Folder¶
You may not configure the import folder to be inside the originals folder, as this will cause a loop by importing already indexed files.
Nested Storage Folder¶
You may not configure the storage folder to be inside the originals folder unless the name starts with a
. to indicate that it is hidden, as this will cause a loop when indexing the thumbnails of already indexed files:
In practice, this is not a problem as long as you follow our documentation and examples. We intend to automatically detect, report, and prevent such issues in the future. Due to the flexible mapping of folders within a Docker container to host directories, this has not yet been implemented.
Symbolic links to files and directories within the originals folder are supported if they are accessible from the environment in which your instance is running. However, you cannot mount a symbolic link as a storage folder or use links within the storage folder.
Session and user management have been reimplemented recently. If you are upgrading from a Development Preview with a build number between 221102-905925b4d and 220901-f493607b0, you will need to run the
photoprism users reset --yes command in a terminal after the upgrade to recreate the new database tables so that they are compatible with the stable version. This will not affect your pictures or albums.
Upgrading from the last stable version should work without any problems. However, if you have already created additional accounts with the previously offered unofficial multi-user support, you will notice that only the main admin account is migrated automatically. Run
photoprism users legacy in a terminal to display the legacy accounts so you can migrate them manually if needed.
Face recognition can be slow (or even crash) on old devices due to insufficient resources.
Like most applications, PhotoPrism has certain requirements and our development process does not include testing on unsupported or unusual hardware.
Asian Faces and Children¶
It is a known issue that children and Asian-looking faces cannot be recognized reliably. Detection without automatic recognition should not be affected by that.
This is because the model we use was trained with North American images, which unfortunately do not include many Asians. The absence of children in the training data comes from the fact that parents do not usually share such images under a public license (and may not have the right to do so).
We will continue to improve our models over time as our resources allow.
Face recognition was developed and tested under the assumption that the background worker runs every 15 minutes, unless the backend is busy with other tasks like indexing. It has not been tested with much longer intervals and is not designed for that.
PhotoPrism's background worker groups new faces by similarity, compares faces with clusters, and optimizes existing clusters as needed. Without these routine tasks, the number of faces to be processed becomes too large. The first and next time the worker runs, it can then cause a heavy server load until all the faces, face clusters, and related pictures have been updated. The longer you wait, the more CPU is required and the longer it takes.
An important reason for the worker to run independently of actual changes in the main instance is that some users change the database content directly or run additional instances, for example for indexing. It is a problem that can be solved, but it takes time. If we were to ignore this and don't run the worker at all times, it could lead to many additional support requests, further reducing the amount of time we can spend on development.
The handling of changes in multiple instances will be improved over time so that the worker can be run less frequently in future releases.
Photos With All Faces Assigned Appear Under “New Faces”¶
This can happen when multiple image files are grouped into a stack, for example, because they were taken at the same place and time, and stacking is enabled.
Secondary images are not searched for faces by default. So the problem is limited to specific cases where you have manually changed the primary image or the images are stacked after detection, which may indicate more fundamental problems, e.g. with the metadata, filenames, or settings.
One possible solution is to change the primary image of a stack to assign faces to the other images in the stack. You can also manually unstack these files and disable stacking in Settings > Library. Note that files that are already stacked are not automatically unstacked when you change the stacking settings, and that Live Photos do not appear in Stacks because they are a special type of media that is always "stacked".
Removing Merged Clusters Fails¶
Under certain conditions, inconsistent face assignments cannot be automatically resolved by the background worker, which can result in an unusually high CPU load when it is running:
- if you use multiple browser tabs or windows for assigning faces and don't wait until saving the changes is complete, the likelihood of this problem increases, especially if you accidentally enter different names for the same face
- another possible cause is running multiple instances (for example, parallel indexing workers started by a scheduler in the background) or modifying database content directly, as this may also lead to inconsistent faces, markers and subjects
- see Faces: Error "Failed removing merged clusters for subject" seems to cause tagging of faces to become slow #2806
Running the following command in a terminal can resolve problems with inconsistent data:
docker compose exec photoprism photoprism faces audit --fix
It can also be helpful to manually check for inconsistent assignments and fix them in the user interface.
Alternatively, you can use the
photoprism faces reset command for a clean start if you haven't invested much time in assigning faces yet.
Advanced users affected by this are welcome to privately provide us with a SQL dump of their subjects, faces, and markers database tables for debugging. Thank you very much!
JPEG: Bad RST Marker¶
This error can occur when decoding JPEG images that contain consecutive 0xFF bytes, e.g. images that have a "glich" (like a few lines of missing image information at the end) or were created with software that inserts them for padding, although this is based on an edge case of the specification and rather uncommon:
- Bug: (invalid JPEG format: bad RST marker) #1673
- image/jpeg: "bad RST marker" error when decoding #40130
- JPEG: Automatically check and repair broken/invalid images #2463
JPEG Size Limit¶
RawTherapee and "heif-convert" cannot limit the resolution of JPEG files when converting files from other formats such as RAW, DNG, HEIC or AVIF. In general, when converting images, the resolution of the generated JPEG files can be limited with the environment variable
PHOTOPRISM_JPEG_SIZE or the CLI parameter
However, this does not work with certain converters because, unlike Darktable, they do not support CLI options to limit JPEG size:
It would probably also hurt indexing performance and image quality if PhotoPrism reduced the size of the generated file after conversion, for example, by using a temporary file.
As a result, this option is ignored when generating JPEG files with these converters. Whether RawTherapee or "heif-convert" are used depends on additional settings such as
Before reporting a bug, please use our Troubleshooting Checklists to determine the cause of your problem. If you have a general question, need help, it could be a local configuration issue, or a misunderstanding in how the software works:
When reporting a problem, always include the software versions you are using and other information about your environment such as browser, browser plugins, operating system, storage type, memory size, and processor.
We kindly ask you not to report bugs via GitHub Issues unless you are certain to have found a fully reproducible and previously unreported issue that must be fixed directly in the app.
Note that all issue subscribers receive an email notification from GitHub whenever a new comment is added, so these should only be used for sharing important information and not for discussions, questions or expressing personal opinions.