当前位置:网站首页>2022-02-28 incluxdb high availability planning

2022-02-28 incluxdb high availability planning

2022-06-12 22:06:00 Blood dragon waving wings

Catalog

Abstract :

The problem is :

suffer TSM Storage engine implementation impacts :

workload :


Abstract :

influxdb Achieve high availability in advance , This article details how to deal with high availability

The problem is :

  1. The newly added data node synchronizes data , There is data written repeatedly , Whether it will cause problems ?
  2. Synchronize data to other data nodes , There is data written out of order , Whether it will cause problems ?
  3. key After hashing , The service is not interrupted when the slot is migrated , How to design ? Reference resources ceph?

suffer TSM Storage engine implementation impacts :

  1. In the writing tsm Before document , It's going to work on the memory cache Sort
  2. wal Write the structure for order
  3. cache by map structure

workload :

  1. Add the concept of data node group under fragmentation , Divide the data node group under a specific partition
  2. modify meta Maintain the data node network structure , Expand the data node group under the partition
  3. modify meta Service Interface , The purpose is to add a data node group
  4. Modify the data node network connection structure , Use data node groups to connect to each other in the cluster
  5. Add and kakfa Interface of interaction
原网站

版权声明
本文为[Blood dragon waving wings]所创,转载请带上原文链接,感谢
https://yzsam.com/2022/02/202202281153382786.html