search

letta-ai / letta

Letta (formerly MemGPT) is a framework for creating LLM services with memory.
https://letta.com
Apache License 2.0
12.96k stars 1.42k forks source link

Multiple deployment of letta sever pointing to same db results in duplicate persona and human entries #2079

Open raolak opened 4 days ago

raolak commented 4 days ago

Describe the bug I have requirement where I need more than one letta service (server per workflow). But when I install the server, seems like duplicate data's being ingested for Human and persona blocks. May be its part of boostraping server. We can make all the name fields of entities with unique constraint, so that we will not allow any duplicate entities creation with same name.

Refer attached screenshot

Please describe your setup

Screenshots Attached ![Uploading Screenshot 2024-11-21 at 9.59.46 AM.png…]()

sarahwooders commented 4 days ago

Why do you need multiple letta servers? The service is designed to support multiple agents. There are probably conflicts between the servers since they are connecting to the same DB.

raolak commented 4 days ago

Hi Sarah,

The discussion is outlined here: GitHub Issue Comment https://github.com/letta-ai/letta/issues/2021#issuecomment-2468817753.

A single server could suffice if the tools didn’t rely on local resources. However, in my case, I embedded Letta alongside my service because the tools need to operate on local resources. For example, tasks like cloning, Docker build and deploy, and other similar use cases require local processes. Additionally, scenarios like analytics may not necessitate executing tools on the Letta server itself, as they depend on interacting with local resources and processes.

Hope my understanding is clear ! And may be there is a better way to do this .

Best regards,

_lnr

On Thu, 21 Nov 2024 at 10:24 AM, Sarah Wooders @.***> wrote:

Why do you need multiple letta servers? The service is designed to support multiple agents. There are probably conflicts between the servers since they are connecting to the same DB.

— Reply to this email directly, view it on GitHub https://github.com/letta-ai/letta/issues/2079#issuecomment-2490077404, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGPZOPKMQY7C55XBGJDSPT2BVRRDAVCNFSM6AAAAABSGCE5W6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOJQGA3TONBQGQ . You are receiving this because you authored the thread.Message ID: @.***>