milvus/tests/python_client/scale
ThreadDao 55e62cc7ff
[skip ci] Update scale data node test (#12879)
Signed-off-by: ThreadDao <yufen.zong@zilliz.com>
2021-12-07 15:53:03 +08:00
..
constants.py [skip ci] Update scale data node test (#12879) 2021-12-07 15:53:03 +08:00
helm_env.py [skip ci] Add default scale release prefix (#9683) 2021-10-11 20:18:29 +08:00
README.md Update scale test README (#7872) 2021-09-14 13:11:54 +08:00
scale_common.py [skip ci] Add deploy default mic in scale (#11931) 2021-11-16 21:25:33 +08:00
test_data_node_scale.py [skip ci] Update scale data node test (#12879) 2021-12-07 15:53:03 +08:00
test_index_node_scale.py [skip ci] Add deploy default mic in scale (#11931) 2021-11-16 21:25:33 +08:00
test_proxy_scale.py [skip ci] Remove unused import of scale proxy (#8793) 2021-09-28 19:30:03 +08:00
test_query_node_scale.py [skip ci] Update scale query node test (#12729) 2021-12-03 20:29:51 +08:00

Scale Tests

Goal

Scale tests are designed to check the scalability of Milvus.

For instance, if the dataNode pod expands from one to two:

  • verify the consistency of existing data

  • verify that the DDL and DML operation is working

Prerequisite

Test Scenarios

Milvus in cluster mode

  • expand / shrink dataNode pod

  • expand / shrink indexNode pod

  • expand / shrink queryNode pod

  • expand / shrink proxy pod

How it works

  • Milvus scales the number of pods in a deployment based on the helm upgrade

  • Scale test decouple the milvus deployment from the test code

  • Each test scenario is carried out along the process:
    deploy milvus -> operate milvus -> scale milvus -> verify milvus

  • Milvus deployment and milvus scaling are designed in helm_env.py

Run

Manually

Run a single test scenario manually(take scale dataNode as instance):

  • update milvus helm chart path
  export MILVUS_CHART_ENV=/your/milvus-helm/charts/milvus  
  • run the commands below:
  cd /milvus/tests/python_client/scale  
  
  pytest test_data_node_scale.py::TestDataNodeScale::test_expand_data_node -v -s  

Nightly

still in planning