Are physical node exceptions migrated to other available resource nodes immediately in the cloud platform? Or is it human intervention? Manually recover the exception node of the cloud host ...?
Automatic migration implementation is not difficult, difficult on how to locate the fault? Physical machine exceptions have n assumptions:
Network failure: Switch port anomaly, broadcast storm ... Will this cause multiple node exceptions? An automatic migration of this type of failure can lead to unexpected risks
Service exception: MQ message backlog, disk full, will cause all nodes to not communicate properly, if this failure to implement automatic migration brings unexpected risk
Software exception: Backend storage node service down, monitor exception, will cause the cloud host can not work, how to locate the VM failure, whether to migrate?
Hardware failure: If locating physical node hardware failure, such as hard disk failure, CPU Crash
This article from "Welcome comments, Welcome to like" blog, please be sure to keep this source http://swq499809608.blog.51cto.com/797714/1726935
How node anomalies in the cloud platform consider migration factors