-
Notifications
You must be signed in to change notification settings - Fork 3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Bug]: [benchmark][cluster] search, query, flush failed in DML & DQL scenario, search error message=incomplete query result #29656
Comments
@xige-16 @yanliang567 |
When the TimeTravel functionality was previously removed, it inadvertently affected the MVCC functionality within the system. This PR aims to reintroduce the internal MVCC functionality as follows: 1. Add MvccTimestamp to the requests of Search/Query and the results of Search internally. 2. When the delegator receives a Query/Search request and there is no MVCC timestamp set in the request, set the delegator's current tsafe as the MVCC timestamp of the request. If the request already has an MVCC timestamp, do not modify it. 3. When the Proxy handles Search and triggers the second phase ReQuery, divide the ReQuery into different shards and pass the MVCC timestamp to the corresponding Query requests. issue: #29656 Signed-off-by: zhenshan.cao <zhenshan.cao@zilliz.com>
When the TimeTravel functionality was previously removed, it inadvertently affected the MVCC functionality within the system. This PR aims to reintroduce the internal MVCC functionality as follows: 1. Add MvccTimestamp to the requests of Search/Query and the results of Search internally. 2. When the delegator receives a Query/Search request and there is no MVCC timestamp set in the request, set the delegator's current tsafe as the MVCC timestamp of the request. If the request already has an MVCC timestamp, do not modify it. 3. When the Proxy handles Search and triggers the second phase ReQuery, divide the ReQuery into different shards and pass the MVCC timestamp to the corresponding Query requests. issue: milvus-io#29656 Signed-off-by: zhenshan.cao <zhenshan.cao@zilliz.com>
When the TimeTravel functionality was previously removed, it inadvertently affected the MVCC functionality within the system. This PR aims to reintroduce the internal MVCC functionality as follows: 1. Add MvccTimestamp to the requests of Search/Query and the results of Search internally. 2. When the delegator receives a Query/Search request and there is no MVCC timestamp set in the request, set the delegator's current tsafe as the MVCC timestamp of the request. If the request already has an MVCC timestamp, do not modify it. 3. When the Proxy handles Search and triggers the second phase ReQuery, divide the ReQuery into different shards and pass the MVCC timestamp to the corresponding Query requests. issue: #29656 pr: #29749 Signed-off-by: zhenshan.cao <zhenshan.cao@zilliz.com>
Occasionally throws error: incomplete query resultargo task:fouramf-multi-vector-f2hrf server:
client logs: test result:
test steps:
|
issue: #29656 /kind bug Signed-off-by: xige-16 <xi.ge@zilliz.com> --------- Signed-off-by: xige-16 <xi.ge@zilliz.com>
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
…ry (#675) 1. Support retry on milvus errors. 2. Retry on ErrInconsistentRequery for search and hybrid search operations. issue: milvus-io/milvus#29656 related milvus pr: milvus-io/milvus#31343 --------- Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
/assign @wangting0128 |
occurred againargo task:inverted-corn-vjvws server:
client pod name: inverted-corn-vjvws-3305228201 test steps:
test result:
|
occurred againargo task:inverted-corn-vjvws server:
client pod name: inverted-corn-vjvws-2315877231 test steps:
test result:
|
Please use the latest version of pymilvus, as the retry logic for this error is only available in the newest version (> 2.4.0rc60). @wangting0128 |
Signed-off-by: Nian Liu <nian.liu@zilliz.com>
verification passedpymilvus version: 2.4.0rc66 Since this fix is on the SDK side, please synchronize each SDK owner to add retry logic simultaneously. notifications: |
#811) Signed-off-by: Nian Liu <nian.liu@zilliz.com>
so , how can i get the pymilvus 2.4.0rc66 installation package or source code ? thanks |
only for test purpose please: https://test.pypi.org/project/pymilvus/2.4.0rc66/ |
thank you but i still get error when i use hybrid_search() in a collection with more than 2 vector field : code 2200, message=retry run out of 75 retry times, message=incomplete query result ......... i do not understand why? It seems to happen with large amounts of data (now my data less then 2 million) |
Could you please provide the version of milvus you are using? |
milvus:2.4.0-rc1 search() function works well, but hybrid_search fails all the time |
i suggest you retry on latest v2.3.13 or 2.4, and if it reproduced to you , please file a new issue with full logs attached. |
does milvus v2.3.13 support multi-vector search? i need this feature in my task. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
Is there an existing issue for this?
Environment
Current Behavior
argo task: fouramf-multi-vector-t6hds
test case: test_inverted_locust_hnsw_ivf_sq8_dml_dql_cluster
server:
client logs:
fouram_log.log
search error message
search_error_message.log
query error message
query_error_message.log
flush error message
flush_error_message.log
Expected Behavior
No response
Steps To Reproduce
Milvus Log
No response
Anything else?
test result:
The text was updated successfully, but these errors were encountered: