Zone Cluster Question

I have a question about losing a node in a zone cluster. Hopefully I can explain enough for someone to answer.

Let's say I have a zone cluster spread over 3 physical nodes. If Node1 crashes, would I be correct in assuming I would lose all zones on Node1?

If I lost Node1 below, would I lose availability of Web1 and Web2? I'm assuming Node2 and Node3 would not pick up operation of Web1 and Web2.

[Edit: This is Solaris 10]

  
 [Node1]      [Node2]      [Node3]
______________________________________


| Web1         Web3          Web5    |
| Web2         Web4          Web6    |

______________________________________

it depends on how the service is configured...

My initial assessment was something another called a cluster, but I'm not sure it's really set up that way logically.

My assumption is that there are multiple hosts running zones with the same application through a load balancer. I'm thinking there's a miscommunication of the cluster terminology.

I don't feel the Solaris setup is really anywhere near something like VMware HA. The applications running have their own cluster as well, but it's at the application level.

Everyone seems to think that a node failure won't affect a zone associated with the node. I don't believe that's true in this setup, but I need to confirm it.

I'm working on getting information from the managed environment, but I'm from a linux/windows background and am trying to wrap my head around Oracle's terminology.

The documents I've read about failover/scalable scenarios somewhat confuse me more since the "how-tos" and documentation really don't explain much from a logical/feature/why perspective. I get a few paragraphs and then the commands to implement something basic. I need to know the "why" before running commands.

All-in-all, I'm aware certain things are possible. From the most basic "cluster" set-up, do you think my assumption would be correct.

EDIT: In addition, do you know any way for me to verify the setup. I'm not currently able to run commands on the hosts, but I would be able to submit a ticket to have the commands run. I'm working on getting access to the systems now, but that has not happened.

an easy way to have all available information about a system is the explorer, which is part of the STB.

Thanks, this looks like a terrific tool. I'll have to take a close look at using this to gain a better understanding of my production environment.