On Mar 22, 2007, at 12:43 PM, Sikich, Frank J. wrote:
Also to note we just implemented z/VM on two CEC's and plan to allow failover between the two machines. I am hoping this will resolve some issues. We are in the infant stages of implementing z/ VM and z/Linux but we are trying to get it right the first time. I have searched and found all the HA doc's and Redbooks but I am looking for any input on this issue.
Make sure everyone agrees what "failover" means. Is it the ability to move a running system image from one node to the other without anyone, including connected users, being able to notice? That's really hard.
Is it the ability to detect failure at one node and bring up the machine hosting the failed services on the other node? That's pretty easy.
If you don't get everyone to agree that "that's what we mean by failover" beforehand, the people who want you to fail will find a reason to tell you why what you're doing isn't failovery enough.
Adam
