Delphix Products

 View Only
  • 1.  Adding SQL Failover Cluster as Target Environment in Delphix

    Posted 06-30-2021 04:22:00 AM

    Hi everyone,

    I am not highly experienced in SQL Server databases and having difficulties regarding to SQL Server Failover Clusters and adding them as Target environments in Delphix.

    Unfortunately Delphix Documentation page regarding this operation is not clear for me so we're having problems when trying to add an SQL Server Failover Cluster as Target environment.

    I wanted to reach out to you for asking some questions:

    1) When adding Cluster Servers as Single Target Environments to Delphix, does both Cluster Nodes need to have a Clustered Disk? 

    2)
    What are the size requirements for this Clustered Disk?

    3)
    Do I have to add 2 completely different Clustered Disks to both Cluster Nodes? Meaning that I will have Disk1 and add this to Cluster Node 1 as Clustered Disk and I need another Clustered Disk 2 that needs to be added to Cluster Node 2?

    4)
    Where do I have to install Delphix Connector in Cluster Node1 and Cluster Node 2 before adding these Nodes to Delphix as Target environments? Is it the location C:\Delphix\DelphixConnector for both Cluster Node 1 and Cluster Node 2?

    Here are the documentation I was able to find and checked out so far:
    1) Not clear enough for me: Adding a SQL Server Failover Cluster Target Environment (delphix.com)
    2) Old documentation: Adding a SQL Target Environment – Made Easy (delphix.com)
    3) Not much related to SQL Server Failover Clusters FAQ: SQL Server AlwaysOn Availability Groups (KBA1603) - Delphix

    Anyone who can share any info about this with me?
    Thanks a lot,





    ------------------------------
    Rahim Cetinel
    Lead DataOps Architect | Delphix Blackbelt
    ------------------------------


  • 2.  RE: Adding SQL Failover Cluster as Target Environment in Delphix

    Posted 06-30-2021 11:58:00 AM

    Hi Rahim,


    I use this as the guide:
    https://docs.delphix.com/docs/datasets/sql-server-environments-and-data-sources/managing-sql-server-environments-and-hosts/adding-a-sql-server-failover-cluster-target-environment

    I will answer the best I can and do reply back for clarification.  I will be out of office tomorrow through Monday but back Tuesday.....

    1. You create a clustered disk for each clustered instance - hence the disk will be ONE disk and you will configure it to be clustered and cluster will be able to manage it.
    I don't have the details on creating this shared disk but sites like this might provide some assistance:
    https://www.sqlshack.com/how-to-add-a-new-disk-to-an-existing-sql-server-failover-clustered-instance/

    That is only an example and is not the defacto solution. Typically, this aspect falls under the storage/windows admin and there might be specific procedures on how they accomplish this.  In essence you create ONE clustered disk and assign it to the cluster. It is added to the clustered instance resource group.

    2.  I don't know that there are size requirements.  It will be used for the mount point of the Delphix VDB iSCSI drives so something like 5 or 10 GB should be more than enough (we store the VDB powershell scripts locally for each VDB - each SCRIPT dir is taking less than 1MB). The storage of VDBs is contained on the Delphix storage array.

    3.  Each clustered disk is assigned a drive letter - you can create more than one if you like but it is created as a disk per Cluster, not per node. It might be easier to create more than one drive per VDB project, or if you run multiple instances on the cluster you can assign a specific drive to one instance.  So like Drive D: on instance 1, Drive E: on instance 2, etc (These are clustered instances not standalone).  You are only creating ONE drive.  The cluster is aware of this when assigned and during failover the theory is the Failover Cluster will provide the drive to the other node where failover occurs.

    4.  You install DelphixConnector on each Node - you can use default location or you can use your own - just make it the same so it is consistent.  Once done you simply discover these as the stand alone nodes.  One very important point - you will also need to provide a non-clustered standalone target environment to act as the cluster proxy/witness.  This also has DelphixConnector installed.  I suggest wherever you install you make it consistent on all the standalones just so it is easy to locate and identify when you need to access the installed location.

    I do hope this makes sense!  Long ago I used instructions from a colleague to setup a cluster and I do know it is difficult but once you have process down you will understand it. Now we use scripts in our Datacenter to perform the tasks.  Let us know if you need further clarification.

    Cheers,
    -Paul



    ------------------------------
    Paul Vero
    Principal Technical Support Engineer
    Delphix
    ------------------------------



  • 3.  RE: Adding SQL Failover Cluster as Target Environment in Delphix

    Posted 07-01-2021 09:07:00 AM
    Dear Paul,

    Thanks a lot for the clarification. Your answers would definitely help.

    BR,

    ------------------------------
    Rahim Cetinel
    Lead DataOps Architect | Delphix Blackbelt
    ------------------------------