none
3 node cluster AG quorum configuration RRS feed

  • Question

  • Hi  All,

    i am going to setup 3 node AG as follows:

    2 node in  DC1 (HA)

    1 node in DC2 (DR)

    Every node will have some active SQL DBs  and some  Passive SQL DBs.

    my requirement is to keep cluster running even after any 2 nodes goes down, should i use fileshare witness quorum configuration? what if i'll not use it?

    Regards

    sachin

     


    SQL Server DBA

    Monday, November 6, 2017 9:07 AM

All replies

  • What version of the Windows Server operating system are you using?

    Edwin M Sarmiento Microsoft Data Platform MVP | Microsoft Certified Master/Solutions Master
    Blog | Twitter | LinkedIn
    Learn SQL Server High Availability and Disaster Recovery


    Monday, November 6, 2017 4:22 PM
    Moderator
  • Thanks Edwin for the response,

    we are going to use windows  server 2016


    SQL Server DBA

    Monday, November 6, 2017 5:40 PM
  • What you want to achieve is not always possible, specially when disaster happens.

    You have 2 new options in windows server 2012 R2

    1. Dynamic witness and

    2. Dynamic quorum.

    Now Dynamic quorum means if nodes go down one by one your cluster could remain online on a Node and Witness which could be either FS or Disk. More details Here.

    If majority of your nodes come down at same time, like a disaster then dynamic quorum would not work.

    Before going ahead do a POC and read more about it, make sure it gives you what you want


    Cheers,

    Shashank

    Please mark this reply as answer if it solved your issue or vote as helpful if it helped so that other forum members can benefit from it

    My TechNet Wiki Articles

    MVP


    Monday, November 6, 2017 6:02 PM
    Moderator
  • It is recommended to have a witness regardless if you are running Windows Server 2012 R2 and higher. However, with your requirement of "keep cluster running even after any 2 nodes going down", this is not possible. Because the possible combination of 2 nodes down out of 3 nodes cannot guarantee that those two nodes will be on different data centers. If those 2 nodes happen to be in DC1, you're out of luck.

    I suggest you redefine your requirement by explicitly defining your recovery objective (RPO/RTO) and service level agreement. Have a process in place to define the action items in case the 2 nodes on DC1 both go down and you need to meet your recovery objectives. It's not that Windows Server 2016 cannot handle this. But technology should not be the basis for providing the solution for your requirement.


    Edwin M Sarmiento Microsoft Data Platform MVP | Microsoft Certified Master/Solutions Master
    Blog | Twitter | LinkedIn
    Learn SQL Server High Availability and Disaster Recovery


    Monday, November 6, 2017 7:35 PM
    Moderator