edward.zeng
edward.zeng
**I've tried the version, the query shard exception has gone.** While I make another query with highlighting, the returned result which matches the query has no 'highlight'. The index setting:...
/assign @LoveEachDay
/assign @XuanYang-cn
From the milvus log, it seems that milvus hangs in `AssignSegmentID`. ``` [2022/09/20 16:32:51.724 +00:00] [INFO] [datacoord/services.go:146] ["handle assign segment request"] [collectionID=436126291330272580] [partitionID=436126291330272581] [channelName=by-dev-rootcoord-dml_1_436126291330272580v1] [count=25017] [isImport=false] ["import task ID"=0] [2022/09/20...
@Sivannnnnn We've enabled data compression for vector data in minio (#15840) which brings 3.3 compression ratio. Here's our test with 100 million data (128 dimension) with index type `ivfsq8` ,...
> and in my environment, dataCoord.enableGarbageCollection = false，  You should enable `enableGarbageCollection` .
> hello @LoveEachDay ,The configurations I use are the default configurations and have not been modified. In addition, I have many machines and there is no case of multiple deployments...
@zhengsonglinge data inside `milvus/rdb_data` cleanup will take some time to iterate all the topics. Could you verify the disk usage now?
@Uzay-G Did you keep the default etcd config in the `docker-compose.yaml`? Here a revised docker-compose yaml which enabled pprof, so that you can check etcd memory usage: ``` version: '3.5'...