标题: HACMP Question [打印本页] 作者: chenpager 时间: 2002-02-13 00:09
提示: 作者被禁止或删除 内容自动屏蔽作者: larryh 时间: 2002-02-14 02:52
Yes, HACMP don\'t swap alias IP address while the corresponding NIC fails\r\nHACMP was primarily designed to deal with node failures, not application failures, so, if you have to monitor application\'s status and choose some actions while it fails, you must define events and scripts by yourself. Anyway, HACMP ES make this job very easy. \r\n\r\nFrom you descrioption, I can conclude these:\r\n\r\nFirst, you must have 12 service IP labels for every instance, because you want to define them in seperate resource groups so that they can be controlled seperately.\r\n\r\nSecond, you don\'t want to make all instances fallover while just some of them fail.\r\n\r\nWhat I have to correct you is that you don\'t have to get 12 standby servers, just 12 resource groups. You can use 1 standby server and 1 standby NIC to backup all 12 application servers, of course, it can take over only one simultaneously. I think this is reasonable, and if you have more NICs on the standby server, it can take over more application servers. 12 shared VGs are necessary.作者: chenpager 时间: 2002-02-14 05:41
提示: 作者被禁止或删除 内容自动屏蔽