For simple functionality, then VCS and MSCS are the same so in a 2-node cluster where node 0 goes offline, both clusters will online the service/cluster group on the remaining node.  

Where VCS really comes into its own is rather than having say six 2-node active-passive clusters, in VCS you can configure one 8-node cluster with 6 active nodes and 2 passive nodes - this uses 4 less nodes and if you do maintenance on an inactive node, then you still have HA, because you have a remaining node to fail to, where as in a 2-node cluster, doing maintenance on the inactive node means you have no HA for this period.  And if you are unlucky enough to have 3 nodes go down then you can configure VCS to fail service groups to a node already runnning a service group, which main mean the service runs more slowly, but it is available.

There are many other advantages of VCS and a lot of these are covered in detail in http://www.symantec.com/content/en/us/enterprise/fact_sheets/b-storage-foundation-ha-windows-6.pdf, 

For installing software, this is pretty much run the installer (UNIX) or setup (windows) on one node and this asks you for nodes in the clusters and software is installed on all nodes.  

At end of installtion it takes you into base configuration to setup heartbeats, cluster users and event notification.  In Windows you can then configure service groups using wizards, but in UNIX, configuring service groups is a manual task which is covered in detail in the VCS agent documentation for the Application/Database you are using.

Mike

So lets suppose i have a 2 node cluster.If my node 0 goes down.Then what actually happens in the background.

What is the importance of service gp?

I have one active and one passive node.

Is it a 2 node cluster?

Yes you have a 2-node cluster.  A service group (or cluster groups) defines a set of components that move together as one unit and can be thought of as sort of a virtual server.  These components (resources) usually consist of storage, virtual IP and application processes, so when node 0 goes down in an active-passive cluster, node 1 detects that it has gone down and onlines the service group on itself - i.e gains control of storage, brings up virtual IP and starts application processes.  In a more complex environment you may have multiple nodes and several service groups on a single node and when the node goes down, the service groups get distributed on the remaining nodes.

Mike

 onlines the service group on itself - i.e gains control of storage, brings up virtual IP and starts application processes

This happens on node 1?i mean node 1 onlines the service gp?

Yes the inactive node 1 onlines the service group.

Mike

ok.Then at what time it goes back to active node.Is it manual or automatic?Can you please tell me the process?

It is a manual operation - you need to switch service to other node, via GUI:

right click on service group and choose "switch" and then choose system (will only be one system with a 2-node)

or via command line:

hagrp -switch group_name -sys sys_name

Switching back causes a brief outage while it is switching which is why it is manual and why it is not recommended - you should leave service running on inactive node as really the 2 nodes in the cluster should be equally capable of running the service, so it shouldn't matter where it is running.

Mike

sys_name here is the node name.?

What is the path of the shared disk location?

How much time it takes from failovering from 1 node to another?Any command which which we can see the time?

sys_name is the node name you are switching the service to.

The path of the shared disk location is whatever you define it to be in VCS resource (example f: in windows or /app1 in UNIX)

The time taken to move the service from one node to another is the time taken to stop processes and deport storage on one node and then import storage and start processes on the target node.  Time taken to import and deport storage depends on number of disks and volumes (and version of storage foundation when you have lots of disks, like over 50), but for each of deport and import, it should be less than 10 seconds for less than 5 disks and 5 volumes.

Mike

If you disable or freeze the service group, then if node 0 goes down, the service group will not online - this is the point of disabling or freezing.  Disabling means VCS stops monitoring the resources where as freezing still monitors the resources and so reports their state (online or offline), but takes no action when resources or nodes go down.

Mike

So in this case failover will not take place?

Yes -  failover will not take place if you disable service group.

Mike

How to check whether failover was successful?What command can be used?or the path from which it can be seen?on node 0 or node 1 ,we can see the logs?

Command:

hastatus -sum

shows you state of service group and which system it is online on - if it is online then failover was succesful.  The command will also show you the status of the systems and any failed resources and any resources or service groups that are frozen or disabled.  

To see the path configured to mount shared storage run:

hares -display -attribute MountPath

The above commands work on both UNIX and Windows

The main log is engine_A.log in:

program files\cluster server\logs

or

/var/VRTSvcs/logs

Mike

OK .Can you please provide me the file with complete path in which i can find the comnplete details of my cluster?