locked
Shared SQL Cluster for Backend - Best Practice/General SfB Advice RRS feed

  • Question

  • Hi,

    Currently we have Lync 2013, single server in an enterprise pool,  single separate SQL 2012 server for backend, single edge server.

    We're about to update to SfB. At the moment I'm planning to get SfB up and running side-by-side and then migrate users over. Again, single server in an enterprise pool (to make expansion easier later on).

    Main reason for not doing an in-place upgrade is that the Lync 2013 servers (FE, BE, Edge) are all Windows Server 2012 (not R2) and we want to move to R2.

    We have recently consolidated a bunch of general SQL servers into a big chunky always-on SQL 2014 cluster.

    Putting performance aside - What is the best practice around locating SfB DBs with other general DBs? Ideally we'd use the shiny new cluster so we don't have to support a standalone SQL server. And of course the additional availability is good.

    As I'm writing this, I'm wondering if I should be moving to standard edition? We have a ~600 users, mainly IM/P but starting to promote and ramp up AV conf. An Enterprise Voice pilot is planned for late 2017. 

    HA/DR is not a requirement at the moment but likely will be within the next 12-18 months. 

    Any advice much appreciated.

    Cheers,

    Rhys


    Thursday, September 22, 2016 9:17 PM

Answers

  • For SQL, you should also post on the SQL forums, but when collocating with an existing cluster I typically like to go with a separate instance (again collocated on the same hardware) where I can control memory consumption of each instance so they don't kill eachother.  I don't like multiple apps tied to a single instance because maintenance of one system can affect the others, and if EV may be in your future that could become a big deal.

    If you don't have EV, and it won't be a pilot until 2017, then standard edition will make your life much easier... :)  You can pilot with identical functionality on standard and always build a new pool later and move people over when HA becomes necessary.


    Please remember, if you see a post that helped you please click "Vote" on the left side of the response, and if it answered your question please click "Mark As Answer". SWC Unified Communications This forum post is based upon my personal experience and does not necessarily reflect the opinion or view of Microsoft, SWC, their employees, or other MVPs.

    • Proposed as answer by Liinus Tuesday, September 27, 2016 5:31 AM
    • Marked as answer by jim-xu Monday, October 10, 2016 8:33 AM
    Thursday, September 22, 2016 9:23 PM
  • It might, to be honest, after countless deployments, I've never actually seen a Lync or Skype server go down outside of a scheduled maintenance.  I've deleted them and perform DR for a test, but never had to in the real world.  As long as you treat it with respect due a phone system (don't install antivirus on the server because some IT guy decided everything needs it), it's super stable.

    Please remember, if you see a post that helped you please click "Vote" on the left side of the response, and if it answered your question please click "Mark As Answer". SWC Unified Communications This forum post is based upon my personal experience and does not necessarily reflect the opinion or view of Microsoft, SWC, their employees, or other MVPs.

    • Proposed as answer by Liinus Tuesday, September 27, 2016 5:31 AM
    • Marked as answer by jim-xu Monday, October 10, 2016 8:33 AM
    Monday, September 26, 2016 8:43 PM

All replies

  • For SQL, you should also post on the SQL forums, but when collocating with an existing cluster I typically like to go with a separate instance (again collocated on the same hardware) where I can control memory consumption of each instance so they don't kill eachother.  I don't like multiple apps tied to a single instance because maintenance of one system can affect the others, and if EV may be in your future that could become a big deal.

    If you don't have EV, and it won't be a pilot until 2017, then standard edition will make your life much easier... :)  You can pilot with identical functionality on standard and always build a new pool later and move people over when HA becomes necessary.


    Please remember, if you see a post that helped you please click "Vote" on the left side of the response, and if it answered your question please click "Mark As Answer". SWC Unified Communications This forum post is based upon my personal experience and does not necessarily reflect the opinion or view of Microsoft, SWC, their employees, or other MVPs.

    • Proposed as answer by Liinus Tuesday, September 27, 2016 5:31 AM
    • Marked as answer by jim-xu Monday, October 10, 2016 8:33 AM
    Thursday, September 22, 2016 9:23 PM
  • Thanks for that. I guess I'll go and lab out the always-on scenario.

     I'm wondering if paired standard edition pools would make a good long term solution. 

    Monday, September 26, 2016 6:35 PM
  • It might, to be honest, after countless deployments, I've never actually seen a Lync or Skype server go down outside of a scheduled maintenance.  I've deleted them and perform DR for a test, but never had to in the real world.  As long as you treat it with respect due a phone system (don't install antivirus on the server because some IT guy decided everything needs it), it's super stable.

    Please remember, if you see a post that helped you please click "Vote" on the left side of the response, and if it answered your question please click "Mark As Answer". SWC Unified Communications This forum post is based upon my personal experience and does not necessarily reflect the opinion or view of Microsoft, SWC, their employees, or other MVPs.

    • Proposed as answer by Liinus Tuesday, September 27, 2016 5:31 AM
    • Marked as answer by jim-xu Monday, October 10, 2016 8:33 AM
    Monday, September 26, 2016 8:43 PM