none
NLB Works only when I stop the host

    Question

  • Hello All,

                I have a two-2 server CAS Array that is NLBed.  I can access individual servers OWA, but when I try to access the NLBED Address, it fails; same is with the NLBED ip address.

    However, when i stop one server in the cluster, it works both with ip address and hostname?  I have tried stopping both servers with the other one active and it works.  This is very strange.  Am i missing something?

    thanks in advance.

    Thursday, February 23, 2012 10:28 AM

All replies

  • Hi,

    Try to validate the state of the NLB, you must see it green and working.

    I assure that you have a DNS record for the casarray. try also to use NLB with multicast mode.

    Regards,


    Best Regards Don't forget to mark it as answer if it helps

    Thursday, February 23, 2012 10:55 AM
  • Thanks for the reply.

    Yes, the host record is there.

    1-  Do you have to do all the configuration from one server?  Because I create cluster from server 1 and then from server 2 i added server 2 into the cluster? would this cause a problem?

    2-  Do I need to enable IP fORWARDING?


    • Edited by Geek Seek Thursday, February 23, 2012 11:27 AM
    Thursday, February 23, 2012 11:22 AM
  • Hi,

    1. No this is not problem.. but check state on the two server

    2. yes, ip forwarding should be enabled.

    you can use this step by step configure : http://info.kraftkennedy.com/blog/bid/102109/

    regards,


    Best Regards Don't forget to mark it as answer if it helps

    Thursday, February 23, 2012 1:49 PM
  • Thanks for the reply.

    1-  Thanks for the answer.

    2-  It is enabled.

    I have followed the article you mentioned with no effect?

    Any suggestions?

    Monday, February 27, 2012 6:05 AM
  • let me rephrase it so that people can understand better

    SERVER 1

    NIC 1:10.0.0.1 (NLB)

    NIC 2:10.0.0.2

    SERVER 2:

    NIC 1:10.0.0.3 (NLB)

    NIC 2:10.0.0.4

    NLB ADDRESS: 10.0.0.5 AGAINST HOSTNAME MAIL.ABC.COM

    All ports are configured on NLB.

    Running in Unicast mode.

    IP forwarding has been enabled on both CAS.

    RPC CAS ARRAY SET TO MAILBOX DATABASES.

    SCRATCHING MY HEAD WHAT AM I MISSING?

    Monday, February 27, 2012 1:08 PM
  • hello, any update on above?
    Tuesday, February 28, 2012 1:05 PM
  • I have followed the article step by step, but still no affect.  As soon as I add the other host, the page cannot be displayed error comes up.
    Tuesday, February 28, 2012 7:52 PM
  • Can two NICs on the same network cause issues?  Should I try with single NIC? OR Multicast Mode? 
    Tuesday, February 28, 2012 8:43 PM
  • Hello All,

                I have a two-2 server CAS Array that is NLBed.  I can access individual servers OWA, but when I try to access the NLBED Address, it fails; same is with the NLBED ip address.

    However, when i stop one server in the cluster, it works both with ip address and hostname?  I have tried stopping both servers with the other one active and it works.  This is very strange.  Am i missing something?

    thanks in advance.

    First are you running the CAS from a VMware?

    http://www.vmware.com/files/pdf/implmenting_ms_network_load_balancing.pdf    

    "NOTE:VMware recommends that you use multicast mode, because unicast mode forces the physical

    switches on the LAN to broadcast all Network Load Balancing traffic to every machine on the LAN."

    htttp://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1006778

    Second are you using any cisco equiptment?  http://www.cisco.com/en/US/products/hw/switches/ps708/products_configuration_example09186a0080a07203.shtml

    Good luck

     

    I wanted to add that when I had this problem I could ONLY connect to the CAS-ARRAY from within the same subnet, so for example I used the Mailbox server and pointed to the CAS-ARRAY and it worked, however if I tried it from my laptop on another subnet it would fail.

    • Edited by GabooN Tuesday, February 28, 2012 9:20 PM My Example
    Tuesday, February 28, 2012 9:01 PM
  • Thanks for the reply.  I am not using VMware or Cisco.

    Let me give you the background again.

    SERVER 1

    NIC 1:10.0.0.1 (NLB)

    NIC 2:10.0.0.2

    SERVER 2:

    NIC 1:10.0.0.3 (NLB)

    NIC 2:10.0.0.4

    NLB ADDRESS: 10.0.0.5 AGAINST HOSTNAME MAIL.ABC.COM

    All ports are configured on NLB.

    Running in Unicast mode.

    IP Forwarding has been enabled on both CAS at NLB NIC.

    RPC CAS Array Set to Mailbox Databases.

    When I create cluster and add first host, everything works fine, but as soon as I add the second host, nothing works.  After that when I stop any of the node, it starts working.  Essentially, cluster only works when one host is enabled.  The mode is unicast.  All port are allowed.  IP Forwarding is enabled on NLB NICs on both servers.

    It has drove me crazy.

    Wednesday, February 29, 2012 8:08 PM
  • Hello All, I would appreciate if someone could comment on above.  Thanks.

    Tuesday, March 06, 2012 7:53 PM
  • Hi Gieek,

    I have exactly the same problem, identical to explain your case. Were you able to resolve it? Someone who has had the same problem and was able to solve?

    thank you very much

    Tuesday, June 04, 2013 8:06 PM