Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Cluster setup

Status
Not open for further replies.

instinct

Programmer
Aug 23, 2003
9
CA
I am interested in clustering and am having trouble finding good sources on how to do it. What I would like to know is how to setup just 2 computers to parallel compute, share memory, and drives. Could someone give me a step by step set up for this or a webpage that has it.
Done with any version of windows/win server or vmware.

Thank you.
 
What I would like to know is how to setup just 2 computers to parallel compute, share memory, and drives."

It really sounds like you want some form of application clustering. The term "Cluster" can mean many different things. MSCS provides failover clustering using a shared storage model. This means that the nodes participating in the cluster all access a shered SCSI bus, however only one node at a time is able to access any given drive. The node controlling the disk resource accomplishes this by using a SCSI reserve command to put a lock on the drive. Should the node fail, the surviving node(s) arbitrate for and take over access to the resources, then start the services.

WLBS uses multiple identical servers that "listen" to the same virtual IP and service requests. Once a client gets a connection to a given node, it will continue to talk to that node until the session ends or times out. This is sometimes called affinity. Hardware load balancing systems like F5, BigIP, or Load Director do essentially the same thing. It's important to note that this is a shared nothing model.

Many applications provide partitioning or clustering at the application level through multiple virtual instances and backend replication. Gnoats would be an example of this. Other applications are written to divide a problem up into smaller chunks and assign the bits to multiple servers which process them in parallel. The SETI screensaver application would be an example.

I think you need to further define the problem in order to find the right solution.

 
I have been donig more research on clustering and my description of what I wanted to do was very vague.

The first step I want to take is to get 2 computers to parallel process in other words combine processors of the 2 different computers.

Can someone help me with that.

 
There isn't a Cluster model for parrallel processing.
Both nodes of the Cluster can run processes at the same time, but they cannot share the processing. An application can only run on one node or the other.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top