I’m experiencing an issue with the Anything-LLM container after restarting it. I installed Anything-LLM via the GUI on TrueNAS, using Ollama as the LLM provider (running locally). I’m using LanceDB as the default database.
Initially, everything functions correctly. However, upon restarting the container, I consistently lose the database connection and the LLM provider configuration. I have to re-enter the provider and re-enter the URI for the search engine each time. Fortunately, previous chat history and user configurations are preserved.
I’m not suspecting a permissions issue, as I haven’t encountered any problems when embedding new documents – they are vectorized and saved to the database correctly.
Has anyone else experienced similar difficulties, or can anyone suggest potential causes for this behavior?
No sorry, unfortunately this was not the solution. Same problem again. Tried to change permissions to user “anything-llm”, as mentioned in the “run-as context”, but nothing changes. In general I don’t think this is a permissions issue, since anythingllm can safe for example the chats or write into the database, tha is located in the dataset actually.
I would like to check if there is a missing mountpoint, but I have no Idea what exactly to look for. The installation was through the TrueNAS GUI, so I don’t have a docker config file. I can check the container through Portainer, but my skills are not that good, to notice whether there is missing something. So I would appreciate any help or advice. Wondering, if I am the only one with this problem, running anythingllm on TrueNAS…
