Tuesday 30 October 2012

CLUSTERING IN WAS


    CLUSTERING IN WAS  

   
Def of CLUSTERING : It is a set of AppServers having same applications Installed,Grouped locally for WORKLOADMANAGEMENT (WLM)
                                 (or)

-----> Grouping of AppServers under a single application / name

-----> Using Clustering we can create a scope and achieved through WORKLOADMANAGEMENT and FAILOVER
                            (OR) 

-----> Every cluster member must have the same configuration and same version.We can say each cluster member is the clone to one another. 

-----> Clustering concept is for High Availability and and Workload Management.



Q)  How to install an application in one cluster member out of multiple members?

A) One cannot deploy application in one cluster member alone in the clustered environment since application which is deployed in one cluster member will be shared across to the other cluster members automatically. So logically it is equivalent, installing application in cluster level.

----> Make the "cluster member weight" where you DONT want the application to be installed as Zero.



There are 2 types of CLUSTERING



1) VERTICAL CLUSTERING : Deployed the application on one machine and we can run that application on our own box(or) machine.

----> In this Vertical Clustering Machine FailOver is not possible.

----> If the machine gets failed we cannot run that application (or) SINGLE POINT OF FAILURE.

---> If one Server fails the other Server will takes care

----> If machine fails nothing can be done.
                            (OR) 

----->  All the JVMs resides on the same physical machine along with the DMGR.

-----> When the JVM failover happens other JVM in that cluster will handle the request as part of the High Availability.

-----> But when the physical machine failover happens we will lose every thing


2) HORIZONTAL CLUSTERING : Grouping of AppServers in which one AppServer should be on one node and Another AppServer should be on another node.

-----> Means if we take two machines one application should be deployed on different machines Remotely (or) one application should run in INDIA and another application should run in AMERICA.

-----> If Machine1 application will be deployed on Machine2 so we have to know the IP address of Machine1 then only it is possible to deployed that application Successfully.

----> Both Machines should have Internet Connection.

----> Here no SINGLE POINT OF FAILURE. 

----> It Supports MACHINE FAILOVER.

                            (OR) 

-----> DMGR and the JVMs resides on the different machines.This will work even when the physical machine failover is happened.


WORKLOADMANAGEMENT : Sharing requests across Multiple Servers.

SCALABILITY            :  can add 2/3 members or add a cluster member to existing member

LOADBALANCING   :  Allocate workload proportionality among available resources.

AVAILABILITY          : System runs if server fails with Clustering.  (or)

---->  Applications are still available if a server fails.



Q)    What is work load manager in WebSphere? What is the default work load management policy?



A) ----> Workload management (WLM) is a WebSphere facility that provides load balancing and affinity between application servers in a WebSphere clustered environment. 

----> WLM is an important fact of performance. WebSphere uses workload management to send requests to alternate members of the cluster. 

----> WebSphere can also be configured to route concurrent requests from a user to the application server that serviced the first request. This is called session affinity and can be used to maintain a users session over concurrent HTTP requests

----> WLM is configurable. The administrator should ensure that each machine or server in the configuration processes  share  the overall client load that is being processed by the system as a whole. 

----> Workload should be spread among machines such that the workload corresponds to the machine processing power.



Q)  What is High Availablity in WebSphere?

A) ----> High Availability means nothing but clustering concept .

 ----->Application should be live for the 365 day for those kind of application they kept in cluster. When the request comes from client it should handle. this concept is called as high availablity

                                        (0R) 

-----> Application availability for the users who are using it. 

------> We can provide the high availability by installing the application in different cluster members. 

-----> In case one of the Application Servers fails other Application server which is in the cluster group will server the request so then the customer request can served quickly.

VERTICAL CLUSTER CONFIGURATION :

           
STEP 1:  Goto Cluster.

STEP2 : Click New.

STEP3: Give Cluster Name.

STEP4 : Prefer LOCAL (By default enable)

STEP5 : Click Next Button.

STEP6: Add Cluster Member 1

STEP7 : Weight (2 Members ---> Equivalent weight)

STEP8 : Add Cluster Member2

STEP9 : Generate unique HTTP ports.

STEP10 : Click Finish Button.

STEP11 : Select Scope.

STEP12 : Controlling Cluster.

STEP13 : Start Cluster Member1

STEP14 : Start Cluster Member2

STEP15 : LogOut from AdminConsole.

STEP16 : Again LogIn to AdminConsole.

STEP17 : See in AdminConsole both Clusters are started.

                         (or)

STEP BY STEP FOR CREATING CLUSTERS.

STEP  1   :    ND - None Installation.

STEP  2   :   Create 3 profiles DMGR , APPSRV1 , CUSTOM.

STEP  3   :  Federation (APPSRV1 , CUSTOM) ------> DMGR.

STEP  4   :  IHS   (WEB SERVER + PLUG-IN  )

STEP  5   :  Map Web Server Instance in DMGR.

STEP  6   :  Cluster Creation  (Node A)  ---->  SRV 2
                                                  (Node B) -----> SRV 1

STEP   7   :   Deployed Application under Cluster or Web Server Scope

STEP   8   :  Ensure PLUG-IN Generation and Propogation.

STEP   9   :  Restarts (SERVER 1  ,  SERVER   2  ,   WEB SERVER)


0 comments:

Post a Comment