locked
App-V High Availability Design RRS feed

  • Question

  • Hi,

    I am in the process of designing a resilient and Scalable App-V infra. The design is generic and should be able to scale to the increase in user count.

    Data store will be SQL Server 2008 which will clustered.

    2 or 3 App-V Management Server across Hardware Load Balancer.

    Applications will be accessed using "Read only Shared Cached feature" which will be setup in a high performance SAN.

    My question is on the App-V Management Server Service, how to go about building reduncdancy for this service, as the IPD guide states " App-V does not offer any automatic fault tolerance for the Management Server service, so it is a single point of failure to be monitored in the system".?

    Can the App-V Management Server Service installed on all available App-v Management Servers so that there is a redundancy in it? What are disadvantages of having the Management Service on more then one App-V Management Servers?

    What are the other impotant points that needs to be addressed in this design?

    Thanks and Appreciate your help.

    Regards,

    Vijay S

     

     

     

    Friday, December 16, 2011 9:35 AM

Answers

  • Both App-V Management servers are independent of each other but use the same data store, so therefore see the same data. The Management Server is stateless - it hold no data (unless the Content share is copied to each Management Server)

    If a server goes down and all requests are handled by the second server, changes are still written to the SQL database. When the first server is brought back up, it will see the new data in the database.



    Twitter: @stealthpuppy | Blog: stealthpuppy.com

    This forum post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

    Please remember to click "Mark as Answer" or "Vote as Helpful" on the post that answers your question (or click "Unmark as Answer" if a marked post does not actually answer your question). This can be beneficial to other community members reading the thread.

    Thursday, December 22, 2011 5:01 PM
    Moderator

All replies

  • If you're placing the App-V Management Server service behind a hardware load-balancer, then you should achieve the resiliency you need.

    See these links for some discussion on configure Citrix NetScalers to support the Management Server:

    The Management Server service should be installed on as many servers as you need for load-balancing and redundancy. The App-V Management Web Service is different however and is only needed to manage the App-V database, so it doesn't need to go on all servers, but it may help simplify the design by keeping all servers configured the same.

    You'll also need to ensure that the Content share is highly available using your SAN and DFS.



    Twitter: @stealthpuppy | Blog: stealthpuppy.com

    This forum post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

    Please remember to click "Mark as Answer" or "Vote as Helpful" on the post that answers your question (or click "Unmark as Answer" if a marked post does not actually answer your question). This can be beneficial to other community members reading the thread.


    Friday, December 16, 2011 10:35 AM
    Moderator
  • Hello,

    Since you are using a read-only shared cache I would focus availability on that specific file and the storage its on.

    I would recommend making the file-server area highly available which hosts the content-folder.

     


    Nicke Källén | The Knack| Twitter: @Znackattack
    Saturday, December 17, 2011 12:32 PM
  • Thanks Aaron and Znack for your reply.

    I have another clarification in the above design.

    If I place the management server service behind behind the load balance  that is installing it along with the app-v management server, then how the resiliency works?

    For example I have two App-V servers and I am installing the management server service in each of the app-v servers. Both the app-v servers are behind the load balancer. Normally the  Management server service will come in to picture, if we are making any changes to the app-V db or for querying AD information.

    When a users access/launches an application, the request is sent to the management server which then relays it to the App-v DB for the necessary permissions. Now if the server which has management server  or the management service is down, how the management server service in the second App-v server will function? If we restore the management server service in the first app-v server, what will happen to the working management server service in the second appp-v server?

    Any information is highly appreciated.

    Regards,

    Vijay S

     

    Wednesday, December 21, 2011 3:40 PM
  • It's the job of the load balancer to ensure the services are available before directing client connects to an available server. If a server goes down, the load-balancer should no longer direct clients to it. When the server is available again, it can then make it available.

    Twitter: @stealthpuppy | Blog: stealthpuppy.com

    This forum post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

    Please remember to click "Mark as Answer" or "Vote as Helpful" on the post that answers your question (or click "Unmark as Answer" if a marked post does not actually answer your question). This can be beneficial to other community members reading the thread.

    Wednesday, December 21, 2011 3:44 PM
    Moderator
  • Thanks for your reply aaron.

    My question is more specific to the App-V Management Server Service and how it works incase of a server down issue. Let me provide you some scenarios:

    1)I have installed App-V Management server in two Servers A and B behind a couple of Hardware Load Balancers. I have also installed the App-V Management Server Service and the management Console in both A and B Servers. During normal operation, I will be accessing the Management Console in Server A to make any changes to the Datastore with the help of management Server Service.

    If Server A goes down  then I will not be able to make any changes to the App-V Datastore(like adding/removing Apps or contacting AD for group information). However since the Hardware load balancer is there, which will automatically redirect the user request for applications to the second server B and the user will continue to access the virtualized apps.

    My question regarding App-V Management Server Service is given below:

    1) In a normal scenario when both the App-V servers are working fine, how the management Service will work when there are two request from two different users which are being directed to the two different servers by the load balancer (incase there is a loading issue)? I would like to know how they are accessing the data store, will it be independent of the other management server service in the other machine?

    2)Can we setup any monitoring for the Management Server Service, so that we get some form of notification when it goes down?

    3)In a two server scenario, when one server goes down the administrators start using the second server for making any modifications to the applications using the management server and the management console in the second server. Once the first server ( which went down initially) is restored, how it will affect the access to the data store? If there are two separate admins who are accessing the app-v admin console in two different servers and trying to modify the same applications, how the app-v handles the conflict as both of them will be using the same service account?

    Any information is highly appreciated.

    Regards,

    Vijay S

    Thursday, December 22, 2011 2:36 PM
  • Both App-V Management servers are independent of each other but use the same data store, so therefore see the same data. The Management Server is stateless - it hold no data (unless the Content share is copied to each Management Server)

    If a server goes down and all requests are handled by the second server, changes are still written to the SQL database. When the first server is brought back up, it will see the new data in the database.



    Twitter: @stealthpuppy | Blog: stealthpuppy.com

    This forum post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

    Please remember to click "Mark as Answer" or "Vote as Helpful" on the post that answers your question (or click "Unmark as Answer" if a marked post does not actually answer your question). This can be beneficial to other community members reading the thread.

    Thursday, December 22, 2011 5:01 PM
    Moderator