diff --git a/site/en/faq/operational_faq.md b/site/en/faq/operational_faq.md index 099567345..e51d0b5b3 100644 --- a/site/en/faq/operational_faq.md +++ b/site/en/faq/operational_faq.md @@ -128,6 +128,7 @@ You receive this warning if the graphics memory required for a request is larger If you use SQLite to manage Metadata, you receive this error message when write requests occur frequently. We recommend using MySQL for Metadata management. See [Manage Metadata with MySQL](data_manage.md). + #### Got an error `Segmentation Fault` from PyMilvus. What shall I do? PyMilvus v1.1.0 allows you to download the latest version of the grpcio library. The server of Milvus v1.x is built on a earlier, customized version of gRPC, which is incompatible with the latest grpcio library. To fix this issue, either upgrade your PyMilvus version from v1.1.0 to v1.1.1 or roll back your grpcio library to 1.37.0: @@ -137,6 +138,11 @@ pip install grpcio==1.37.0 pip install grpcio-tools==1.37.0 ``` +#### Can I specify the timeout for `Milvus()`? + +Yes, PyMilvus v1.1.2 supports specifying the server connection timeout. + + #### Still have questions? You can: diff --git a/site/en/faq/product_faq.md b/site/en/faq/product_faq.md index c5fbdd2fc..ebef39940 100644 --- a/site/en/faq/product_faq.md +++ b/site/en/faq/product_faq.md @@ -192,6 +192,19 @@ See [Index Types](index.md) for more information. {{fragments/faq_max_vector_dimension.md}} +#### Why does Milvus set an upper limit of 256 MB for the data that can be inserted at a time? + +Data inserted to Milvus is first written into memory. This limit is to avoid over-occupation of memory resources. + +#### Milvus merges multi-thread query requests in one batch. How to cancel this mechanism? + +You can add the following section to the configuration file **server_config.yaml**: + +```yaml +engine_config: + search_combine_nq: 1 +``` + #### Still have questions? You can: diff --git a/site/en/home/home.json b/site/en/home/home.json index 06c30ded2..d2b4c2534 100644 --- a/site/en/home/home.json +++ b/site/en/home/home.json @@ -29,7 +29,7 @@ "order": -1, "title": "Introduction", "desc": [ - "Milvus is a cloud-native, open-source vector database built to manage embedding vectors generated by machine learning models and neural networks. It extends the capabilities of best-in-class approximate nearest neighbor (ANN) search liraries (e.g. Faiss, NMSLIB, Annoy) and features on-demand scalability, unified Lambda structure, and high availability. The goal of Milvus is to simplify unstructured data management and provide a consistent experience across different deployement environments.", + "Milvus is a cloud-native, open-source vector database built to manage embedding vectors generated by machine learning models and neural networks. It extends the capabilities of best-in-class approximate nearest neighbor (ANN) search libraries (e.g. Faiss, NMSLIB, Annoy) and features on-demand scalability, unified Lambda structure, and high availability. The goal of Milvus is to simplify unstructured data management and provide a consistent experience across different deployement environments.", "Milvus is widely used in scenarios such as computer vision, natural language processing, computational chemistry, personalized recommender systems, and more. It has been adopted by over 1,000 organizations worldwide including iQiyi, Kingsoft, Tokopedia, and Trend Micro. More than 2,300 developers have joined the Milvus open-source community on GitHub, Slack, mailing lists, and Wechat. ", "Milvus was released under the open-source Apache License 2.0 in October 2019, and its source code was made available on GitHub. In June 2021, Milvus graduated from the LF AI & Data Foundation's incubator program." ] @@ -111,7 +111,7 @@ "link": "https://zilliz.com/blog/Vector-Similarity-Search-Hides-in-Plain-View", "title": "Vector Similarity Search Hides in Plain View", "abstract": "Discover what vector similarity search is, its various applications, and the public resources making artificial intelligence more accessible than ever.", - "imgSrc": "https://zillizstorage.blob.core.windows.net/zilliz-assets/zilliz-assets/assets/small_plainview_703d8497ca.jpg?38953.77500005998" + "imgSrc": "https://zilliz-cms.s3.us-west-2.amazonaws.com/small_plainview_703d8497ca.jpg?38953.77500005998" } ] } diff --git a/site/zh-CN/faq/operational_faq.md b/site/zh-CN/faq/operational_faq.md index d0fe711cc..c72ab6d90 100644 --- a/site/zh-CN/faq/operational_faq.md +++ b/site/zh-CN/faq/operational_faq.md @@ -35,7 +35,7 @@ Milvus 和服务端配置文件的版本不对应。 #### 为什么 Milvus 查询召回率一直不理想? -在调用 SDK 进行向量搜索时,可以增大函数中 `nprobe` 参数的值。值越大,结果越精确,但耗时也越久。详见 [如何设置 Milvus 客户端参数](https://www.milvus.io/cn/blogs/2020-2-16-api-setting.md)。 +在调用 SDK 进行向量搜索时,可以增大函数中 `nprobe` 参数的值。值越大,结果越精确,但耗时也越久。详见 [如何设置 Milvus 客户端参数](https://www.milvus.io/cn/blog/2020-2-16-api-setting.md)。 #### 为什么更新过的设置没有生效? @@ -122,6 +122,7 @@ Milvus 是以 Docker 镜像形式发行的,是可以离线部署的: 如果元数据管理用的是 SQLite,在有数据频繁写入的情况下会出现该错误。建议将 SQLite 更换为 MySQL。如何更换请参考文档 [使用 MySQL 管理元数据](data_manage.md)。 + #### PyMilvus 返回 `Segmentation Fault` 错误,如何解决? PyMilvus v1.1.0 允许下载最新版本的 grpcio library,而 Milvus v1.x 的服务端则基于早期的定制版本 gRPC 搭建,该版本 gRPC 与最新版本 grpcio library 不兼容。解决此问题需将 PyMilvus 升级至 v1.1.1 或将 grpcio library 回滚至 1.37.0: @@ -133,6 +134,10 @@ pip install grpcio-tools==1.37.0 +#### 是否可以指定 `Milvus()` 接口的超时时间? + + +PyMilvus v1.1.2 已支持指定服务器连接超时时间。 #### 仍有问题没有得到解答? diff --git a/site/zh-CN/faq/performance_faq.md b/site/zh-CN/faq/performance_faq.md index 1c25d6452..742f18a18 100644 --- a/site/zh-CN/faq/performance_faq.md +++ b/site/zh-CN/faq/performance_faq.md @@ -62,7 +62,7 @@ id: performance_faq.md 如果向集合中导入数据的频率不高,建议将 `index_file_size` 的值设为 1024 MB 或者 2048 MB。如果后续会持续地向集合中导入增量数据,为了避免查询时未建立索引的数据文件过大,建议这种情况下将该值设置为 256 MB 或者 512 MB。 -可参阅 [如何设置 Milvus 客户端参数](https://www.milvus.io/cn/blogs/2020-2-16-api-setting.md)。 +可参阅 [如何设置 Milvus 客户端参数](https://www.milvus.io/cn/blog/2020-2-16-api-setting.md)。 #### Milvus 的导入性能如何? diff --git a/site/zh-CN/faq/product_faq.md b/site/zh-CN/faq/product_faq.md index 7e18ed6b4..805be6ec1 100644 --- a/site/zh-CN/faq/product_faq.md +++ b/site/zh-CN/faq/product_faq.md @@ -108,7 +108,7 @@ ID 类型是非负的 64 位整型。 #### Milvus 中如何选择向量索引的类型? -索引需要根据具体场景和需求去选择。详见 [如何选择索引类型](https://milvus.io/cn/blogs/2019-12-03-select-index.md)。 +索引需要根据具体场景和需求去选择。详见 [如何选择索引类型](https://milvus.io/cn/blog/2019-12-03-select-index.md)。 #### Milvus 可以在同一个集合中的不同分区上建立不同索引吗? @@ -132,7 +132,7 @@ ID 类型是非负的 64 位整型。 所以当总的向量条数约等于 `nlist` 时,两者的计算量相当,性能也差不多。而随着向量条数达到 `nlist` 的 2 倍、3 倍、n 倍之后,IVF_FLAT 的优势就越来越大。 -可参阅 [如何选择索引类型](https://milvus.io/cn/blogs/2019-12-03-select-index.md)。 +可参阅 [如何选择索引类型](https://milvus.io/cn/blog/2019-12-03-select-index.md)。 #### 创建索引立即查询,为什么内存会突然增长? @@ -196,6 +196,18 @@ collection 数量没有限制。每个 collection 内的 partition 总数不能 #### Milvus 支持的向量维度的最大值是多少? {{fragments/faq_max_vector_dimension.md}} +#### 为什么 Milvus 对单次可插入的数据大小设置了 256MB 的上限? + +插入 Milvus 中的数据会被先写入内存。 这个限制是为了避免插入操作过度占用内存资源。 + +#### Milvus 会批量合并多线程查询请求。 如何取消这个机制? + +你可以将以下内容添加至配置文件 **server_config.yaml** 中: + +```yaml +engine_config: + search_combine_nq: 1 +``` #### 仍有问题没有得到解答? diff --git a/site/zh-CN/fragments/faq_flat_ivfflat.md b/site/zh-CN/fragments/faq_flat_ivfflat.md index f0323c274..e9a464dc8 100644 --- a/site/zh-CN/fragments/faq_flat_ivfflat.md +++ b/site/zh-CN/fragments/faq_flat_ivfflat.md @@ -5,5 +5,5 @@
所以当总的向量条数约等于 nlist
时,两者的计算量相当,性能也差不多。而随着向量条数达到 nlist
的 2 倍、3 倍、n 倍之后,IVF_FLAT 的优势就越来越大。
如果向集合中导入数据的频率不高,建议将 index_file_size
的值设为 1024 MB 或者 2048 MB。如果后续会持续地向集合中导入增量数据,为了避免查询时未建立索引的数据文件过大,建议这种情况下将该值设置为 256 MB 或者 512 MB。
nprobe
参数的值。值越大,结果越精确,但耗时也越久。详见 如何设置 Milvus 客户端参数。
\ No newline at end of file
+在调用 SDK 进行向量搜索时,可以增大函数中 nprobe
参数的值。值越大,结果越精确,但耗时也越久。详见 如何设置 Milvus 客户端参数。
\ No newline at end of file
diff --git a/site/zh-CN/fragments/faq_set_nlist.md b/site/zh-CN/fragments/faq_set_nlist.md
index f8db28764..03341447f 100644
--- a/site/zh-CN/fragments/faq_set_nlist.md
+++ b/site/zh-CN/fragments/faq_set_nlist.md
@@ -1 +1 @@
-该值需要根据具体的使用情况去设置。详见 性能调优 > 索引 和 如何设置 Milvus 参数。
\ No newline at end of file
+该值需要根据具体的使用情况去设置。详见 性能调优 > 索引 和 如何设置 Milvus 参数。
\ No newline at end of file
diff --git a/site/zh-CN/home/home.json b/site/zh-CN/home/home.json
index 9f0c673bc..533b0a6f9 100644
--- a/site/zh-CN/home/home.json
+++ b/site/zh-CN/home/home.json
@@ -109,7 +109,7 @@
"link": "https://zilliz.com/blog/Vector-Similarity-Search-Hides-in-Plain-View",
"title": "Vector Similarity Search Hides in Plain View",
"abstract": "Discover what vector similarity search is, its various applications, and the public resources making artificial intelligence more accessible than ever.",
- "imgSrc": "https://zillizstorage.blob.core.windows.net/zilliz-assets/zilliz-assets/assets/small_plainview_703d8497ca.jpg?38953.77500005998"
+ "imgSrc": "https://zilliz-cms.s3.us-west-2.amazonaws.com/small_plainview_703d8497ca.jpg?38953.77500005998"
}
]
}
diff --git a/site/zh-CN/milvus_basics/index.md b/site/zh-CN/milvus_basics/index.md
index 5ffe710a1..db034fdce 100644
--- a/site/zh-CN/milvus_basics/index.md
+++ b/site/zh-CN/milvus_basics/index.md
@@ -297,7 +297,7 @@ Annoy(Approximate Nearest Neighbors Oh Yeah)是一种用超平面把高维
## 选择索引
-- 若要为你的使用场景选择合适的索引,请参阅 [如何选择索引类型](https://milvus.io/cn/blogs/2019-12-03-select-index.md)。
+- 若要为你的使用场景选择合适的索引,请参阅 [如何选择索引类型](https://milvus.io/cn/blog/2019-12-03-select-index.md)。
- 关于索引和向量距离计算方法的选择,请访问 [距离计算方式](metric.md)。
diff --git a/site/zh-CN/reference/data_manage.md b/site/zh-CN/reference/data_manage.md
index 2108a37bc..aabf3590c 100644
--- a/site/zh-CN/reference/data_manage.md
+++ b/site/zh-CN/reference/data_manage.md
@@ -68,8 +68,8 @@ Milvus 在 CentOS 系统中不支持 MySQL 8.0 或更高版本。
从数据导入,数据存储到数据查询和调度,请参阅我们的博客深入了解 Milvus 数据管理方案。
-- [数据管理策略](https://www.milvus.io/cn/blogs/2019-11-08-data-management.md)
-- [数据文件清理机制的改进](https://www.milvus.io/cn/blogs/2019-12-18-datafile-cleanup.md)
-- [查看元数据](https://www.milvus.io/cn/blogs/2019-12-24-view-metadata.md)
-- [元数据表的字段](https://www.milvus.io/cn/blogs/2019-12-27-meta-table.md)
-- [如何通过元数据管理数据文件](https://www.milvus.io/cn/blogs/2020-01-09-milvus-meta.md)
+- [数据管理策略](https://www.milvus.io/cn/blog/2019-11-08-data-management.md)
+- [数据文件清理机制的改进](https://www.milvus.io/cn/blog/2019-12-18-datafile-cleanup.md)
+- [查看元数据](https://www.milvus.io/cn/blog/2019-12-24-view-metadata.md)
+- [元数据表的字段](https://www.milvus.io/cn/blog/2019-12-27-meta-table.md)
+- [如何通过元数据管理数据文件](https://www.milvus.io/cn/blog/2020-01-09-milvus-meta.md)
diff --git a/site/zh-CN/release_notes.md b/site/zh-CN/release_notes.md
index 21dce2e50..59186167f 100644
--- a/site/zh-CN/release_notes.md
+++ b/site/zh-CN/release_notes.md
@@ -598,7 +598,7 @@ id: release_notes.md
- **HNSW 索引的支持**
- 新增了对 HNSW 索引类型的支持。关于 HNSW 的详细介绍请参考[向量索引算法 HNSW 和 NSG 的比较](https://milvus.io/cn/blogs/2020-01-16-hnsw-nsg-comparison.md)。[#853](https://github.com/milvus-io/milvus/issues/853)
+ 新增了对 HNSW 索引类型的支持。关于 HNSW 的详细介绍请参考[向量索引算法 HNSW 和 NSG 的比较](https://milvus.io/cn/blog/2020-01-16-hnsw-nsg-comparison.md)。[#853](https://github.com/milvus-io/milvus/issues/853)
- **Jaccard/Hamming/Tanimoto 距离的支持**