【文件属性】:
文件名称:优先级策略-PRIMECLUSTER介绍及日常维护
文件大小:2.54MB
文件格式:PPT
更新时间:2021-04-22 18:28:08
富士通 PRIMECLUSTER 维护
优先级策略
共享存储
Node B
Start Job A
Node A
Job A
Stop Job A
Job B
Stop Job B
Problem!
不但提供了更高的可用性,还可以为各种提供不同的优先级
This is a variation of the previous scenario. Here the the main job is too big to share resources so it is failed-over tov a machine where the applications can be stopped. Typically this is a development machine. So for the time of the failure all development work stops to allow all resource be available for the real-time application.
Development work can either be restarted on the failed node once it is recovered, or the application swapped back.