[Bug]: Stuck in search and the answer is cut.

[fire717]

Is there an existing issue for the same bug?

• [X] I have checked the existing issues.

Branch name

main

Commit ID

8d667d5abdbd185e60140b22150851aa8e6a3ebe

Other environment information

a6000 48g
ubuntu22
docker

Actual behavior

1. When add some files and try to chatting, it stuck with "搜索中" ，and you cannot see the vertical line of the mouse cursor flickering in the end (under normal circumstances, it will flash and then start answering)
2. The model didn't answer a few sentences before it became "......由于长度的原因，回答被截断了，要继续吗“，evenif I change the MAX TOKEN from512 to2048 in the model setting page.And I am sure the token model has answered not even more than 100 token.

Expected behavior

No response

Steps to reproduce

Origin code with the env list before. Add dozens of documents

Additional information

No response

[KevinHuSh]

For problem 1, I bleived it stucked in embedding or reranking since it's slowly running on CPU. Another option is to check ES status, like how much CPU or RAM it used. For 2, disable max_token, just toggle out.

[skytodmoon]

mark， I met the same problem with dev version on 2024-09-11