Best Practices to Handle 50k partitions Per Cluster (Compaction/Indexing追不上的问题) #38996
Unanswered
xiaobingxia-at
asked this question in
Q&A and General discussion
Replies: 3 comments 1 reply
-
stats任务,也就是sort,是在2.5刚刚加入Milvus的特性。 |
Beta Was this translation helpful? Give feedback.
1 reply
-
你要去看一下 get IndexNode slots failed 失败的原因是什么。 |
Beta Was this translation helpful? Give feedback.
0 replies
-
对
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
需要建立大量partition,每个partition小量数据。compaction index一直无法追上。近期把compaction、index code以及log完整读了一遍。梳理了一下大概流程:
继续插入数据:
结束。
所以关于compaction和indexing node的实际工作量,如果是一小时内插入数据到了1000个partition中:
请问以上分析是否准确。
我的问题是:
我目前的理解是,如果单个data node和index node无法并行化,我需要开大量的小体积的data node/index node。而不是开少量大体积的data node/index node。不然就会被卡主。
Beta Was this translation helpful? Give feedback.
All reactions