I am encountering issues starting the Milvus services (rootcoord, datacoord, querycoord, indexcoord). The services continuously restart with errors indicating that permissions for the public/default namespace are denied. Below are the details:

I encountered an issue where Milvus repeatedly logs errors about failing to load segments. The logs indicate that the server keeps trying to load segments but fails with the error not supported at /workspace/source/internal/core/src/storage/DataCodec.cpp:98.

I am getting errors like “too busy disk” and Milvus is crashing every few minutes.

I am trying to use milvus for image similarity search involving 100 million+ embeddings. We have a different column-oriented distributed db table outside of milvus to annotate hundreds of metas for a given image like latitude, longitude, # of pedestrians, etc.

I am using pymilvus version: 2.4.4

We have a milvus collection of about 351M rows on databricks. We are hosting our data on a single ec2 instance standalone milvus container (r5.8xlarge). Our collection has 380 partitions and only have an index on our sparse float vector column.

We’ve been using Langchain and Milvus for a chatbot project and it has been running fine. Recently we noticed that if Milvus is left overnight with a 2009 item collection, the queries fail with “failed to search: out of range in json: ef(10) should be larger than k(20))”.

we are planning to do some 2 step ANN search with 2 kinds of embeddings; word embedding, multimodal embedding.
The scenario is described as the following image.

This is the original directory:

I am looking for an option to export milvus data in ver 2.4 (sort of backup), and then migrate it to another machine. According to the doc, there is such option for migrating from ver 1,x.
Is it possible to perform milvus-migration export from ver 2.4?