如何减轻coordinator的运行压力? #39056
Unanswered
xiaobingxia-at
asked this question in
Q&A and General discussion
Replies: 2 comments 3 replies
-
另外,我看到每隔半小时query node CPU使用率拉高时,log里有大量loading segment, releasing segment的信息。 |
Beta Was this translation helpful? Give feedback.
2 replies
-
可能要首先确定一下这个半个小时一次的是什么任务? |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
cluster 2.5.2, 600k个partition,每个paritition 3500行。有个数据导入script在不停导入数据。每次导入500行数据到一个partition,所以他一小时会碰到上万个partition(产生上万个segment)
有发现如下问题:
根绝我对源代码的理解,应该是coordinator被阻塞瓶颈住了。。。如果错了请指正。
我现在唯一能想到的是,我的segment maxlife是半小时。所以每隔半个小时会flush数据一次。请问还有什么其他可能的原因?
另外,我想优化coordinator的工作,由于我有大量的partition,我想避免coordinator一遍遍频繁的轮询所有segment,我在看以下几个选项:
对于以下选项,是否可以拉高:
3. queryCoord.checkSegmentInterval: 1000
4. checkChannelInterval: 1000
5. checkBalanceInterval: 300000 # check balance every 300 seconds
6. checkIndexInterval: 10000 # check index every 10 seconds
7. dataCoord.syncSegmentsInterval: 300
Beta Was this translation helpful? Give feedback.
All reactions