zilliztech / milvus_cli Goto Github PK

connect db error use milvus_cli0.4.0

connect -uri 192.168.xx.xx:19530 -t root:xxxx
message=I1legal uri: [192.168.xx.xx:19530]，expected form "http[s]://[user :[email protected]
connect to MiTvus error!<ConnectionConfigException: (code=1,
:12345>

使用python 代码没有问题， 在milvus_cli上报错

I installed milvus_cli and readline in RHEL 8. and milvus_cli core dumps at any command.

[root@st2]# milvus_cli



  __  __ _ _                    ____ _     ___
 |  \/  (_) |_   ___   _ ___   / ___| |   |_ _|
 | |\/| | | \ \ / / | | / __| | |   | |    | |
 | |  | | | |\ V /| |_| \__ \ | |___| |___ | |
 |_|  |_|_|_| \_/  \__,_|___/  \____|_____|___|

Milvus cli version: 0.4.0
Pymilvus version: 2.3.0

Learn more: https://github.com/zilliztech/milvus_cli.


milvus_cli > connect
free(): invalid pointer
Aborted (core dumped)

[root@st2]# milvus_cli


  __  __ _ _                    ____ _     ___
 |  \/  (_) |_   ___   _ ___   / ___| |   |_ _|
 | |\/| | | \ \ / / | | / __| | |   | |    | |
 | |  | | | |\ V /| |_| \__ \ | |___| |___ | |
 |_|  |_|_|_| \_/  \__,_|___/  \____|_____|___|

Milvus cli version: 0.4.0
Pymilvus version: 2.3.0

Learn more: https://github.com/zilliztech/milvus_cli.


milvus_cli > help
free(): invalid pointer
Aborted (core dumped)

pip show pymilvus
Name: pymilvus
Version: 2.3.0
Summary: Python Sdk for Milvus
Home-page:
Author:
Author-email: Milvus Team <[email protected]>
License:
Location: /root/jupyter/virtualenv/slack/lib/python3.10/site-packages
Requires: environs, grpcio, pandas, protobuf, ujson
Required-by: milvus-cli

Milvus version: dev version.
Cli: v0.2.0

1. connect milvus
2. query on already loaded collection
3. throw Error: <BaseException: (code=5, message=collection has already been loaded, and the number of replicas 1 is not same as the request's 0. Should release first then reload with the new number of replicas)>

query on unloaded collection, every working fine.

我的版本如下：

• python 3.8.17
• pymilvus 2.3.0
• milvus_cli 0.4.0
  当我使用 milvus_cli 连接到 milvus 后，执行 describe collection -c test_collection ，命令行报错 Error: No such command 'describe'.

I just want to test Milvus on GPU server with milvus_cli. But there is no any GPU_IVF_FLAT option to create index. The output is like that: Milvus cli version: 0.4.0
Pymilvus version: 2.3.0
Learn more: https://github.com/zilliztech/milvus_cli.
milvus_cli >
milvus_cli > create index
Collection name (docs_general, car): car
The name of the field to create an index for (id, vector, color, brand): vector
Index name: idx_vct
Index type (FLAT, IVF_FLAT, IVF_SQ8, IVF_PQ, RNSG, HNSW, ANNOY, AUTOINDEX, DISKANN, ) []:

安装milvus_cli，就会自动卸载之前安装的2.2.9版的pymilvus， 给我安装一个低版本的pymilvus-2.2.1, 导致代码在导入数据时报找不到from milvus import db, 需要重新安装pymilvus?, 希望zilliz工程师能尽快修复版本库包问题

When will the creation of scalar indexes be supported with milvus_cli?

cli failed to start on RHEL 8

./milvus_cli-v0.3.3-Linux
[3800578] Error loading Python lib '/tmp/_MEIzrq14X/libpython3.8.so.1.0': dlopen: /lib64/libm.so.6: version `GLIBC_2.35' not found (required by /tmp/_MEIzrq14X/libpython3.8.so.1.0)

python --version
Python 3.10.12

I tried with python 3.8, failed at the same error.

./milvus_cli-v0.3.3-Linux
[3803912] Error loading Python lib '/tmp/_MEIbq8zhU/libpython3.8.so.1.0': dlopen: /lib64/libm.so.6: version `GLIBC_2.35' not found (required by /tmp/_MEIbq8zhU/libpython3.8.so.1.0)

python --version
Python 3.8.16

I've learned from the Get started with Milvus_CLI and the Milvus_CLI Command Reference, but the command like the import has been removed it seems and still in the docs.

as a macos user, I like to manage all my software using homebrew. Could you add a formula to add the milvus CLI so that it could be installed with:

brew install milvus-cli

Command "show loading_progress" errors anyway no matter how I use this command:


After some look up, I found that milvus_cli passes the partition parameter of tuple type to pymilvus, but pymilvus only accepts the partition parameter of list type, resulting in the above error.
I think there are two solutions to this issuse:

1. Implement a callback function in "show loading_progress" which transform the partition parameter of tuple type to list type.
2. Modify pymilvus to make pymilvus support partition_arrary parameter input of tuple type.

I personally prefer solution 1 because it this has the least impact on the whole Milvus project, and I've already implemented this.
Please let me know what you think.

as picture show. when i run 'show collection -c $collection_name'.i got different response. question collection schema not have json filed. openai_amc_question have json field like 'FieldSchema(name='question_meta', dtype=DataType.JSON),'

please tell me what happened?

Milvus cannot use the dataType of varchar when creating a collection, indicating the need for varchar. However, using varchar will result in an error

command line as follows
milvus_cli > create collection -c test_commond -f id:INT64: -f enterprise_id:VarChar: -f vector:FLOAT_VECTOR:1536 -p id -d test_commond
Error!
Invalid field data type, should be one of ['BOOL', 'INT8', 'INT16', 'INT32', 'INT64', 'FLOAT', 'DOUBLE', 'STRING', 'BINARY_VECTOR', 'FLOAT_VECTOR']
milvus_cli >
milvus_cli > create collection -c test_commond -f id:INT64: -f enterprise_id:VARCHAR: -f vector:FLOAT_VECTOR:1536 -p id -d test_commond
Error!
Invalid field data type, should be one of ['BOOL', 'INT8', 'INT16', 'INT32', 'INT64', 'FLOAT', 'DOUBLE', 'STRING', 'BINARY_VECTOR', 'FLOAT_VECTOR']
milvus_cli >
milvus_cli > create collection -c test_commond -f id:INT64: -f enterprise_id:STRING: -f vector:FLOAT_VECTOR:1536 -p id -d test_commond
RPC error: [create_collection], <MilvusException: (code=1, message=string data type not supported yet, please use VarChar type instead)>, <Time:{'RPC start': '2023-04-20 15:12:31.273735', 'RPC error': '2023-04-20 15:12:31.274766'}>
Error occurred!
<MilvusException: (code=1, message=string data type not supported yet, please use VarChar type instead)>

As usually, we never want to change password by write some python codes and execute it, and authorization is very important in current world, why not just provide it in cli? or maybe I can provide a PR to milvus_cli project to implement this feature

The current iteration of the Milvus CLI lacks Transport Layer Security (TLS) capability, leaving communications unencrypted despite supporting authentication. This enhancement request aims to fortify the CLI by implementing TLS support for secure data transmission between the CLI and the server.
@nameczz

Current milvus_cli depends on python >=3.8.5 to run, while most server image in production (like RHEL<=7 or ubantu <=16) won't have python at such high version. Some slimed or old images may only have python2 or even no python installed.

I'm recently try to build a rpm binary release for RHEL7, and I mean to enclude a milvus_cli for the user to check the service after installation. But I give up for there're so many dependency to install and there're dependency conflicts that could drive our user crazy.

Now that the go SDK is available, maybe we can use it to build a binary client without dependencies.

Although the number of entities is not 0, it always outputs 0.

Milvus cli version: 0.3.0
Pymilvus version: 2.1.0

Getting errors during startup, due to this: minio/minio-py#1378.
I had to manually specify minio==7.2.0 to make it work.

and how to use UsingDatabase、ListDatabases、DropDatabase commands

• Add Dockerfile
• Add workflow to build & push docker image to docker hub when taged

When I run milvus_cli.exe on Windows 10：

it show this error:
C:\Users\yu.wang\Downloads>milvus_cli.exe
Traceback (most recent call last):
File "scripts\milvus_cli.py", line 1, in
ImportError: attempted relative import with no known parent package
[17988] Failed to execute script 'milvus_cli' due to unhandled exception!

How can I solve this problem?