Qdrant
Replies: 1 comment 6 replies
-
|
You can check the shards layout in the cluster by checking the
|
Beta Was this translation helpful? Give feedback.
-
Is there any logs? Or what kind of problem you noticed? |
Beta Was this translation helpful? Give feedback.
-
|
grpc interface raise error: I didn't collect the docker logs, so this is the last logs, maybe I should write it to the disk: |
Beta Was this translation helpful? Give feedback.
-
|
it looks like shard just timeouted all requests 🤔 |
Beta Was this translation helpful? Give feedback.
-
|
this timeouted state will last for a long time, and cannot recover. what I can do is restart docker process of this node. |
Beta Was this translation helpful? Give feedback.
-
|
so if one shard' 6335 timeout for some times, it will be tagged as unavailabe, and the collection won't work? but in fact the shard is ok. |
Beta Was this translation helpful? Give feedback.
-
hello me again!
follow my last discussion.
we have 0.15billion vectors of 768d. 3 nodes have 4cpu, 32GRam, 400G high performance ssd.
we follow the instruction of partition, it works. we write all our data into qdrant. we try some searches, it's ok.
we query qdrant with very low qps. but now one node doesn't work again.
docker logs as below:
its ports of 6334 and 6335 cannot reach.
only the memory reach the limit of docker ---24g.
the information of my collection is as below:
{ "result": { "status": "green", "optimizer_status": "ok", "vectors_count": 160881120, "indexed_vectors_count": 160875833, "points_count": 160881120, "segments_count": 623, "config": { "params": { "vectors": { "size": 768, "distance": "Cosine", "on_disk": true }, "shard_number": 6, "replication_factor": 1, "write_consistency_factor": 1, "on_disk_payload": true }, "hnsw_config": { "m": 0, "ef_construct": 100, "full_scan_threshold": 10000, "max_indexing_threads": 0, "on_disk": true, "payload_m": 16 }, "optimizer_config": { "deleted_threshold": 0.2, "vacuum_min_vector_number": 1000, "default_segment_number": 0, "max_segment_size": null, "memmap_threshold": 20000, "indexing_threshold": 20000, "flush_interval_sec": 5, "max_optimization_threads": 1 }, "wal_config": { "wal_capacity_mb": 1024, "wal_segments_ahead": 0 }, "quantization_config": null }, "payload_schema": { "groupid": { "data_type": "integer", "points": 160881120 } } }, "status": "ok", "time": 0.003013 }and the info of cluster is :
{ "result": { "status": "enabled", "peer_id": 110393092236890, "peers": { "8183754743245770": { "uri": "http://10.0.0.48:6335/" }, "8410915895960002": { "uri": "http://10.0.0.47:6335/" }, "110393092236890": { "uri": "http://10.0.0.49:6335/" } }, "raft_info": { "term": 806, "commit": 89, "pending_operations": 0, "leader": 8410915895960002, "role": "Follower", "is_voter": true }, "consensus_thread_status": { "consensus_thread_status": "working", "last_update": "2023-06-16T03:01:40.822875688Z" }, "message_send_failures": { "http://10.0.0.48:6335/": { "count": 35, "latest_error": "Error in closure supplied to transport channel pool: status: DeadlineExceeded, message: \"Timeout 1000ms reached for uri: http://10.0.0.48:6335/\", details: [], metadata: MetadataMap { headers: {} }" } } }, "status": "ok", "time": 0.000027242 }Edited by @timvisee for readability
Beta Was this translation helpful? Give feedback.
All reactions