[Bug]: load collection with synchrization does not fully load all the segments even the API returns successfully

[yanliang567]

Is there an existing issue for this?

• I have searched the existing issues

Environment

- Milvus version: master-20250717-ee439545-amd64
- Deployment mode(standalone or cluster): cluster
- MQ type(rocksmq, pulsar or kafka):    with wp and sn
- SDK version(e.g. pymilvus v2.0.0rc2): pymilvus 2.6.0rc162

Current Behavior

if search immediately after load() successfully, some previous released data was not in the search results.

Expected Behavior

after load the collection again, the search results should be same as the first time load the collection. But in the example above, id 379(belongs to partition_1) does not in the res5.

Steps To Reproduce

run test_milvus_client_search_v2_new.py::TestSearchV2Independent::test_search_after_partition_release 
run link: https://jenkins.milvus.io:18080/blue/rest/organizations/jenkins/pipelines/Milvus%20HA%20CI/branches/PR-43330/runs/6/nodes/64/steps/110/log/?start=0

Milvus Log

No response

Anything else?

No response

[yanliang567]

/assign @bigsheeper
/unassign

[yanliang567]

1. if I add time.sleep(10) after 2nd load(), the issue does not reproduce any more.
2. Moreover, it does not reproduce on Milvus v2.5.14

[bigsheeper]

Jenkins link is invalid,
https://jenkins.milvus.io:18080/blue/rest/organizations/jenkins/pipelines/Milvus%20HA%20CI/branches/PR-43330/runs/6/nodes/64/steps/110/log/?start=0

[bigsheeper]

I ran several tests using the latest master branch and couldn’t reproduce the issue. @yanliang567

[bigsheeper]

/assign @weiliu1031