windows cluster Active/Active and Active Passive

In Windows cluster Active /Passive, one node will take ownership of the Resources the other node will stay in Stand by.
In Active/Active Mode will the resources ownership be split between both nodes or how does this work ?

Thank you
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

David Johnson, CD, MVPOwnerCommented:
Active/Passive is predictable fail over when the primary fails.
Active/Active configuration enables both controllers to be active doubling your performance during normal operation. However, during a failover scenario performance will a single controller performance.
jskfanAuthor Commented:
how do you distribute resources In Active/Active, Manually or the nodes will figure it out automatically ?
David Johnson, CD, MVPOwnerCommented:
how do you distribute resources In Active/Active, Manually or the nodes will figure it out automatically ? you can't as each is a mirror of the other and the next available will be used. Just to confirm we are talking about file clustering and not some other clustering i.e. SQL/Web Server
SolarWinds® IP Control Bundle (IPCB)

Combines SolarWinds IP Address Manager and User Device Tracker to help detect IP conflicts, quickly identify affected systems, and help your team take near instantaneous action. Help improve visibility and enhance reliability with SolarWinds IP Control Bundle.

jskfanAuthor Commented:
So even in SQL cluster...Active Nodes are mirrored..I mean the same resources on each node and both are owners of resources
Ryan McCauleyEnterprise Analytics ManagerCommented:
My experience is with SQL Server clustering, but I think the same concepts still apply for any function controlled by a Windows Server Failover Cluster - Active/Active doesn't mean you get double the performance or that there are multiple nodes serving client requests, but rather means that multiple nodes in your cluster normally own things. In an "Active/Passive" scenario, all services are normally hosted on a single node, and in the case of a failure of that node, everything modes to your standby node. When you have an "Active/Active" cluster, it just means that one node owns some things and the other owns the rest - in the case of a failure of either node, all services move to the operational node.

If you've clustered a service (like SQL Server), then only one node can be running the service at any given time, and it's easy to see how it works - you're active/passive by nature of their only being one cluster resource group. However, if you've got two instances, you've got a choice - "Active/Passive" (meaning one node runs both SQL Server instances) or "Active/Active" (meaning one node runs each instance). People normally see this as "load balancing", but it's important to be clear that it's not - you're not levelling the load between the nodes or sharing ownership - you're just running a separate service on each node, and all client requests for each service go directly to the relevant node.

As I understand it, a file share set up via Windows Failover Clustering works the same way - at any time, only a single server is handling all client requests - one or more nodes can be standing by as passive nodes, ready to take over control if there's a failure, but one node is doing all the work. You can somewhat mitigate this by having multiple shares and one is "Active" on each server, but all requests for the share go to a specific node - the workload isn't "shared" or "load-balanced" in any way. This is different from how DFS handles file shares - in that situation, multiple hosts can be set up to handle requests and synchronize files between themselves in a "multi-master" configuration, and this feature does not rely on Windows Server Failover Clustering in any way.

Again, this is MSSQL-centric, but here's the best explanation of how "active/active" is handled that I've ever come across:
jskfanAuthor Commented:
Active/Passive is clear..

Active/Active..for instance sql or exchange cluster. If I understand they are mirrored in matter of resources.
Node 1 will have resource1,resource2,resource3,..............resourceX
Node 2 will have the same: resource1,resource2,resource3,..............resource

when a query will come up, then will it go to Node1 or Node2?
When Node1 will go down , will Node 2 service the same requests as Node1?
David Johnson, CD, MVPOwnerCommented:
when a query will come up, then will it go to Node1 or Node2?
When Node1 will go down , will Node 2 service the same requests as Node1?

1. Either
2. Yes
Ryan McCauleyEnterprise Analytics ManagerCommented:
As I mentioned near the end of the second paragraph on my previous answer (I realize it's a bit long-winded), you're not load-balancing when you have an Active/Active cluster - the services are running on only one node at a time (either A or B) and not on both. When a client submits a request, it will go to the server that owns that service. If that server fails, ownership of those services/resources will go to the other node in the cluster and it will begin handling client requests for those services.

So while both nodes will be configured such that they could own resource1, resource2, etc, only one node will own a resource at any given time and it will service all client requests related to that service.
jskfanAuthor Commented:
Ryan McCauley

Sorry it is not clear yet..
I have use Active/Passive in the past.. Resources are Owned by Active Node, when it goes down the Ownership will failover to the Passive node which will become the Active ....That's clear.

When you set up an Active/Active for instance for SQL or Exchange, If I understand Node1 will own the same(Mirrored) Resources as Node2. Correct ? OR the Resources are different in Node1 and Node2 ?

If Mirrored then , based on what criteria client requests will go to Node1 or Node2 ?

If one node crashes then if I understand there is no failover if resources are mirrored , there will only be a burden on one node, since it will have to service all the requests by itself.

You see what I am trying to understand
David Johnson, CD, MVPOwnerCommented:
active/active means that requests are load balanced between the 2
active/passive requests are handled by active node replicated to passive node
Ryan McCauleyEnterprise Analytics ManagerCommented:
The best visual explanation of how Windows clustering works is something I saw on Brent Ozar's site:

Windows Clustering of other services may have different behavior and I'm only intimately familiar with SQL Server, but my understanding is that Windows Clustering always works that way - one node is active and handles all client requests for a service, and the other node is in standby mode, ready to take control of the resources (storage and any services) if the first node fails so that it can handle new client requests.

The confusing comes in when you have multiple services running on a cluster - in that case, you can have an "Active/Active" cluster, which means that both servers are in control of certain resources and are handling client connections. However, each server owns a different resource, and a resource can be owned by only one node at a time. If I have two instances of SQL Server on my cluster, I can choose to set them up as either:

 - Active/Active - one instance of MSSQL running on each node. If one node fails, the other node is now running both instances. However, connections for Instance1 always go to the node that owns that instance, not to both.
 - Active/Passive - both instances on MSSQL run on a single node, and if that node fails, they both move to the other node and become active.

There are services that can be load balanced - like DFS file sharing - and there may be "clustered" services where multiple nodes can handle client requests (MSSQL AlwaysOn gets into this territory), but traditional failover clustering doesn't work that way.

If I'm incorrect or unclear, I would welcome a better explanation to help clear things up, as there seems to be quite a bit of confusion about clustering in general.
David Johnson, CD, MVPOwnerCommented:
there is a difference between windows and SQL clustering
jskfanAuthor Commented:
To make it simple, assuming there is only one instance of SQL in both servers... and we need to configure Active/Active cluster...
would cluster resources be owned by both nodes ? (you said NO)
So there is no Load balancing between Nodes, if one node fails the other will take over the resources that have been owned by the failed node.?
David Johnson, CD, MVPOwnerCommented:
SQL clustering means that only 1 sql server can write to the database it will be the master
Some people use the term active/active to indicate they have an instance of SQL Server running on both nodes.  That is, 2 separate SQL Server instances with separate storage and databases.  I term this a multi-instance cluster because each instance could be running on either node at any given time.  It could be active/passive (both active on node 1, node 2 idle), active/active (one instance active on each node) or passive/active (both active on node 2, node 1 idle).

The problem with that terminology is that it really isn't clear which node is hosting which instance
. SQL server unfortunately doesn't support RAC clustering like Oracle
David Johnson, CD, MVPOwnerCommented:
To be clear are you referring to SQL or file resources they operate differently and the difference seems to be confusing you.. For File/Server(s) you can use active/active but for SQL clustering I'd stay with active/passive.  At least this way you know what is going on at a specific point in time
jskfanAuthor Commented:
I might have come to understand what David is saying about SQL active /active, it will make sense to have active/active only if you have 2 or more sql instances running on each node.

***the way I understand instance,  it runs separate databases from other instances.

let's say both Nodes Have SQLinstance1 and SQLinstance2.
Active/Active means for instance, node1 one is servicing requests for  SQLinstance1 and node2 servicing requests for SQLinstance2, however if one Node goes down the other will service requests for both SLQinstances..

There is no failover...

When there is just one instance of SQL server, then it makes sense to use Active/passive  since only one node can access SQLinstance at a time.

Am I close to understanding it ?
David Johnson, CD, MVPOwnerCommented:
jskfanAuthor Commented:
so fail over of ownership  is automatic ?
Ryan McCauleyEnterprise Analytics ManagerCommented:
Ownership moves to the node that's currently active - if the active node goes offline (or the clustered service fails), Windows Failover Clustering Services will transfer control of that service to the other node and bring the service online there.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
jskfanAuthor Commented:
Thank you Guys
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Windows Server 2008

From novice to tech pro — start learning today.