Potato

Results 12 comments of Potato

state tranfer 的定义应该是直接传内存磁盘里的数据来保证不同机器间的一致性吧,可以参考[6.824 课程第四讲 vmware fault tolerant 第 8-11 分钟](https://www.bilibili.com/video/BV1R7411t71W?p=4)的讲授内容。 mysql cluster 基于 binlog 的主从同步方式实际上也是基于 RSM 吧,即假设应用相同顺序的日志能够保证不同节点上的 mysql 实例达到相同状态?

> State Tranfer 的核心思想是“当状态发生变化后,应在主、从状态同步一致之后再接受新的外部输入” 这个定义有可参考的文献吗?我查了一些文献,感觉学术界对此概念的定义似乎更多的是“当状态发生变化后,主、从状态同步一致的方式”这一出发点来说的,比如直接传内存的数据,或者类似于 raft 的 state machine 的 snapshot 等等,总之都是直接传输状态(应用层状态或者硬件状态均可)而不是需要重放的 operations。 比如我找到一篇 2006 年 OSDI 的论文 [HQ Replication: A Hybrid Quorum Protocol for Byzantine Fault Tolerance](https://www.usenix.org/legacy/events/osdi06/tech/full_papers/cowling/cowling.pdf),其中对 state transfer...

感觉这个概念还是有一定争议。比如对于 vm-ft 这篇论文来说,其主要思想是从硬件层次传输 CPU 指令,中断等来使得 primary 和 backup 达到一致,论文其中的 Output Rule 规定了 primary 只能在得到 backup 的 ack 之后才能返回结果给客户端,否则永远不响应客户端的请求。这样的实现和 MySQL 的 Fully Synchronous Replication 思想基本是一致的,但是 Robert Morris 教授在课程中却将 vm-ft 描述为一个 RSM...

按照周老师引用[维基](https://en.wikipedia.org/wiki/State_machine_replication#State_Transfer)的原话来说吧。 > When a new replica is made available or an old replica is restarted, it must be brought up to the current State before processing Inputs (see Joining). Logically,...

知乎某阿里工程师:https://zhuanlan.zhihu.com/p/151234708?utm_source=qq

Hi, I noticed that you are writing a request size of more than 200M, which is actually too large. We suggest reducing the batch size to reduce the memory pressure...

Hi, this bug is caused by our internal code handling hostname with `-`. We will fix it in the next version. If you're deploying 1.0 for your current environment, could...

Hi, this issue have been fixed in this [PR](https://github.com/apache/iotdb/pull/8468), which will be released in 1.0.1~

Hi, chris , we have added a off level to close system metric, you could take a look for this pr :https://github.com/apache/iotdb/pull/11530

cc @SteveYurongSu BTW, please notice the code smell.