none
DHCP scope options to support failover for IP phones in multi-pool environment RRS feed

  • Question

  • We have Polycom CX600 phones with Lync 2013 CU3 in a site resilient mutiple pool setup.   If we failover our primary site to the secondary site everything works but the phones wont register.  

    How do I configure the DHCP scope options to have an option to register with the backup pool.  

    Thursday, August 21, 2014 3:56 AM

All replies

  • So the scenario I am trying to combat is if the switches in the primary office are still online but the Lync Servers and or the AudioCodes PBX goes down then the phones need to re-register to the backup pool at the DR site.  If the switches are down at the primary site then we aren't using phones anyway, we would have to use USB headsets for everyone or built in laptop speakers. 

    The way its configured now is the switch uses LLDP with a bootp helper pointing to our DHCP server at the primary site.  I am wondering if I can configure the switch to have an additional DHCP bootp helper to point to our DR sites DHCP server.  Obviously I need to create the same scope settings there.  The DR site can already talk to the VOIP vlan and ping it based on all the routing we did so I think that would be ok.  

    The other unknown i have is if I configure a second bootp server on the Juniper switch what is preventing phones from not going there until a failover event occurs, I would almost have to manually force the scope at the primary site to go offline.  Or manually configure the secondary helper when an outage occurs so its not ideal.  


    • Edited by pslager Thursday, August 21, 2014 1:31 PM
    Thursday, August 21, 2014 1:26 PM
  • Hi,

    You may want to check DHCP failover.

    DHCP failover in Windows Server® 2012 is a new feature that enables two Microsoft DHCP servers to share service availability information with each other, providing DHCP high availability.

    All scope information is shared between the two DHCP servers, including active leases.

    You can refer to the link below:

    http://technet.microsoft.com/en-us/library/dn338983.aspx#cluster

    Best Regards,

    Lisa Zheng


    Lisa Zheng
    TechNet Community Support

    Tuesday, August 26, 2014 2:10 AM
    Moderator
  • Thanks I am already running DHCP Failover on Windows 2012.  
    Tuesday, August 26, 2014 12:48 PM
  • Is the issue that provisioned phones lose connection and do not failover to the backup registrar, or that you cannot provision a new phone during the outage?

    The DHCP 43/120 options are only used for the initial PIN Authentication process, so once a Lync Phone Edition device has successfully been provisioned with a user then future registration attempt (e..g power cycle) will utilize cached credentials and simply perform a legacy SRV/A record lookup for the Lync Pool and use the cached client certificate for TLS-DSK authentication.

    Thus the DHCP 43/120 values should not impact the registration of an already provisioned phone, only the ability to sign a user into an unprovisioned phone during an outage of the server referenced in the 43/120 options.


    Jeff Schertz | Microsoft Solutions Architect - Polycom | Lync MVP

    Friday, September 5, 2014 12:15 PM
    Moderator
  • Thanks for the reply basically after a failover the phones are stuck on trying to connect basically a user has to manually click next through the phone to sign in again.  I am doing another DR test in 3 weeks and will be physically onsite this time so I can see exactly what happens.  
    Friday, September 5, 2014 2:32 PM
  • Make sure that current firmware is installed on these phones as pool failover support was resolved ages ago.

    Jeff Schertz | Microsoft Solutions Architect - Polycom | Lync MVP

    Friday, September 5, 2014 3:24 PM
    Moderator
  • Jeff - we recently had a similar issue as is described here.  A power issue in one datacenter caused us to fail over a Lync 2013 pool to it's backup registrar / paired pool in a separate datacenter.  Most Polycom VVX phones reconnected to the failover pool as expected, but some did not.  The resolution was to update DHCP option 43 to point the WebServerFqdn to the failover pool name.  Our analysis was that these phones had expired client certificates, and were unable to sign in bc they attempted to connect to the cert provisioning service URL provided by DHCP option 43, which pointed to the pool name that was down.  We are thinking that the best way to resolve this might be to use a DNS name in the WebServerFqdn option that is GeoDNS with monitoring of the pool for failover (provided by an F5 GTM), so that if a pool was down all requests would be automatically directed to the other pool. 

    Have you seen someone use this type of solution before?  Do you agree with the analysis of the initial issue?

    Tuesday, January 27, 2015 7:39 PM
  • Did you find a solution to this?
    Saturday, June 9, 2018 12:40 PM
  • I realise this was some years ago for you, but did you find a solution to the option 43 pool issue?

    Thanks, Finn.

    Friday, September 20, 2019 12:15 PM