将服务器添加到现有群集后的sstables数量过多 [英] Huge number of sstables after adding server to existing cluster

问题描述

我们正在运行测试cassandra集群的8个节点
运行在单个DC使用Simple snitch和DateTieredCompactionStrategy
Cassandra版本2.1.3
在添加新节点（第9）到我们看到的集群后新加入的服务器上的sstables数量大致等于集群中所有服务器上所有sstables的总和。

we are running test cassandra cluster of 8 nodes running in single DC using Simple snitch and DateTieredCompactionStrategy Cassandra Version 2.1.3 after adding new node(9th) to the cluster we see that number of sstables on newly joined server roughly equals to sum of all sstables on all servers in the cluster. and that number is huge as tens of thousands of sstables on newly added server.

Q1：这是我们应该期望发生的事情吗？

Q1:is that what we should expect to happen?

此外，新添加的服务器似乎没有重载，基本上没有挂起/计划的压缩，但sstables的数量不减少。

Furthermore newly added server seems isn't overloaded, basically there are no pending/scheduled compactions but the number of sstables isn't decreasing.

Q2：可能是不减少sstables数量的原因？

Q2:what could be the reason of not reducing number of sstables?

Q3：我们需要做什么以减少每个服务器的sstables数量？

Q3:what we need to do to reduce number of sstables per server?

感谢您的帮助

推荐答案

这是已知（意外，但不是不足为奇）到DTCS的设计。如果max_sstable_age_days低于ttl截断，则每次都会发生。当从集群中删除节点时，也会发生这种情况。

This is known (unexpected, but not unsurprising) behavior due to the design of DTCS. It will happen, every time, if max_sstable_age_days is below the ttl cutoff. It will also happen when you remove a node from the cluster.

如果我的演讲被接受，我计划在C * Summit 2015讨论这个问题，我已经在 https://issues.apache.org/jira/browse/CASSANDRA-9666

I'm planning on discussing this at C* Summit 2015 if my talk is accepted, and I've proposed an alternative approach to DTCS in https://issues.apache.org/jira/browse/CASSANDRA-9666

这篇关于将服务器添加到现有群集后的sstables数量过多的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！