System and Network Management

 View Only
  • 1.  Network Region Failover Testing

    Posted 10-30-2017 06:14 AM

    We have an issue that the Avaya backbone engineers are having an issue solving so I'm hoping someone out here can assist.

    For years now we have always tested new site install failover by disabling the IP Network Region on the Active Core Server.

    The phones at the remote location will roll within seconds to the ESS.

    Then we disable the NR on the ESS and the phones will roll to the local LSP.

    This works perfect everytime in one of our core environments. But in the other two the phones never drop from the core when disabling the NR.

    The local gateway drops and registers to the ESS or LSP and the local LSP does the same. But the phones stay registered to the Core and create a split registration scenario.

    The phone then have no dialtone and cannot make or receive calls.

    We've checked all of the settings between the cores and they are all the same. We also cannot turn on the Force Phones option because we have two failover points for every Network Region.

    As a test I also removed all inter region connectivity from a remote NR and the disable NR still fails.

    We can pull the network down at the local site and the phones will register the LSP at that point but we don't have people at every site that can do that work. That also doesn't allow us to test failover to the Core ESS.

    Any thoughts would be appreciated.

     

     

     

     

     



  • 2.  Network Region Failover Testing

    Posted 10-30-2017 07:00 AM
    Hi Greg

    If the backbone guys are on it then I'm uncertain I can personally add
    much. That said however I've assumed the server software load and handset
    firmware are all the same. Also reading between the lines are these h323
    handsets?

    If the local handsets are not registering to the LSP then on the face of it
    Id assume they cannot see the local DSP resource. Thats the route Id take.
    Id also confirm the station page has the correct location assigned and that
    the handset ip-addresses are correctly administered in the ip-network-map.
    Also ensure page 3 on the ip-network-region for the LSP is correctly
    administered, comparing against a working example. Im sure you have checked
    all this, especially if the backbone guys are looking at it.

    Im keen to know what the issue is now, keep us posted.

    On 30 October 2017 at 14:13, Greg Terry <iaug-sys-net@lists.iaug.org> wrote:

    > We have an issue that the Avaya backbone engineers are having an issue
    > solving so I'm hoping someone out here can assist.
    >
    > For years now we have always tested new site install failover by disabling
    > the IP Network Region on the Active Core Server.
    >
    > The phones at the remote location will roll within seconds to the ESS.
    >
    > Then we disable the NR on the ESS and the phones will roll to the local
    > LSP.
    >
    > This works perfect everytime in one of our core environments. But in the
    > other two the phones never drop from the core when disabling the NR.
    >
    > The local gateway drops and registers to the ESS or LSP and the local LSP
    > does the same. But the phones stay registered to the Core and create a
    > split registration scenario.
    >
    > The phone then have no dialtone and cannot make or receive calls.
    >
    > We've checked all of the settings between the cores and they are all the
    > same. We also cannot turn on the Force Phones option because we have two
    > failover points for every Network Region.
    >
    > As a test I also removed all inter region connectivity from a remote NR
    > and the disable NR still fails.
    >
    > We can pull the network down at the local site and the phones will
    > register the LSP at that point but we don't have people at every site that
    > can do that work. That also doesn't allow us to test failover to the Core
    > ESS.
    >
    > Any thoughts would be appreciated.
    >
    >
    >
    >
    >
    >
    >
    >
    >
    >
    >
    > -----End Original Message-----
    >


  • 3.  RE: Network Region Failover Testing

    Posted 10-30-2017 07:06 AM

    Jim thanks for the reply. Yes these are H.323. The problem is when we disable the physical network the phones register to the local LSP fine so it can see the DSP resources.

    Also, if there were no resources available I would expect the phones to go into a discover mode but they just sit there registered to the core.

    Everything matches from a configuration standpoint even the gateway firmware as we keep a standard config setup for all of our locations.

    It's a stumper for sure. Backbone guys are really struggling with this one.

     



  • 4.  RE: Network Region Failover Testing

    Posted 10-30-2017 07:19 AM

    Greg,

    What CM version are you running and are you using the standard original Network Regions (1-250) or are you using the new Stub Regions that came out in 6.3 (at least that is when we saw them).

    We have been told after a CM interchange and our phones not failing over and coming back in to service that the Stub Regions have a bug and we have to upgrade our CMs to 6.3.16 to resolve it. 

    Not sure this applies to you but thought I'd throw it out.



  • 5.  RE: Network Region Failover Testing

    Posted 10-30-2017 07:25 AM

    Thanks for that info John.

    We are on 6.3.13. We had to roll back to 13 because of a bug in 14 that impacted something else we have setup. The NRs are in the lower ranges so that shouldn't impact us. Also, the Core that works properly is on the same load so I'm not expecting it to be a firmware issue but you never know.



  • 6.  RE: Network Region Failover Testing

    Posted 01-03-2018 07:03 AM

    I am unsure that I fully understand your environment... but as long as the ip addresses on the endpoints are in the appropriate regions as defined in the ip-network-map they should follow the appropriate behavior.. when you make registration on that NR disabled they can no longer register to PE or CLANs controlled by that controller... so they will follow the alternative gatekeeper list in their network region..    for grins I would take a look in the phone to see what the alternative gatekeeper list contains..  you will need to have snmp enabled and configured and can use a tool like GetIF to view that information...



  • 7.  RE: Network Region Failover Testing

    Posted 01-04-2018 10:29 AM

    Have you used any SNMP tools to examine the recovery list that is inside the phones?  This could give you a clue.

    Tom Lynn

    Principal Engineer

    Nordstrom, Inc,

    tom.lynn@nordstrom.com