Re: Load Balanced TS Servers and Unicast Flooding

From: Neplitude (nicola_at_iafrica.com)
Date: 03/25/05 

Date: Fri, 25 Mar 2005 18:39:28 +0200

Hi Harrison

The problem you have at hand I have experienced many a time. The best thing
to do is the following:

1. Remove each on of your tse servers from the cluster, either by going
through the network load balancing properties of each nic on each server or
by going into the NLB manager and removing each host accordingly.

2. Verify that your cluster ip address has been removed from your ip
settings. Sometimes what happens is that you may remove a client from a
cluster but the virtual ip cluster address is still assigned to the nic.
This could be the possible reason for you having an invalid address
(question mark) message on the interface.

3. You have two network cards so ideally the best solution would be to use
unicast. Assign the one NIC as your network link for user connectivity. Use
the other network card for cluster communication. Just remember to assign a
different ip address ranges for each nic. So for Example, NIC1 used for user
connectivity (Ip addy = company addressing) and NIC2 which is used for the
cluster communication/configuration could be something like 10.0.0.1
upwards. Just remember to change the interface metric of each nic. The
reason for doing this is that you want NIC1 to take prescendence over NIC2.

5.One you have completed the above configuration changes on all your
servers. Start adding them back to the NLB Cluster. Remember to make the
specific changes that are necessary. For example when you start adding the
servers into the nlb cluster make sure you choose NIC2 on each server, cause
that will be the network card to which each server within the cluster will
be communicating with..

Finally just remember to make the necessary port assignment (3389 to 3389),
set your affinity to none (only if using session directory) if not set it to
single.

If you require any more help just post here :)

Cheers for now
Hope that helped!

Regards
Nep

"Harrison Midkiff" <HMidkiff@aviinc.com> wrote in message
news:eWc15RUMFHA.2748@TK2MSFTNGP09.phx.gbl...
> Rebecca:
>
> Thanks for your reply.
>
> I reviewed all the info you posted for me. I'm not sure how I should
> configure my WLBS? My 2 servers are plugged into a new Cisco Catalyst
> 3560 Gigabit switch. My servers are 2 HP ProLiant 360's both with dual
> NICs. When I read the documentation it mentioned something about using 2
> NIC's with a hub. Ultimately I'd like to both servers to operate on this
> switch for best through put. Would you have any recommendation on exactly
> how to configure that?
>
> I made my NIC changes last night on a maintenance window. Unfortunately
> after experiencing problems I put everything back to what it was. I think
> I did reboot the server and the NIC continued to have a red question mark
> on it.
>
> Any advice you have would be greatly appreciated.
>
> Harrison Midkiff
>
>
> ""Rebecca Chen [MSFT]"" <v-rebc@online.microsoft.com> wrote in message
> news:90GVYMTMFHA.1264@TK2MSFTNGXA03.phx.gbl...
>> Hi Harrison,
>>
>> Unicast flooding is expected if the cluster hosts are attached to a
>> switch.
>> In Unicast mode, traffic addressed to the cluster is flooded to all ports
>> on the switch's virtual LAN (VLAN). You can refer to the following
>> article
>> to configure the server when using Unicast:
>>
>> http://www.microsoft.com/resources/documentation/Windows/2000/server/reskit/
>> en-us/Default.asp?url=/resources/documentation/windows/2000/server/reskit/en
>> -us/distrib/dsdc_cls_ZACZ.asp
>>
>> When you use multicase, please refer to the following article to
>> configure
>> the network:
>> 193602 Configuration options for WLBS hosts connected to layer 2 switches
>> http://support.microsoft.com/?id=193602
>>
>> For more information about NLB, refer to the following two FAQs in NLB
>> FAQ:
>> http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technologies/
>> clustering/nlbfaq.mspx
>>
>> Q. What is the Difference Between the Multicast and Unicast Modes of
>> Operation? Which One Should I Use in My NLB Installation?
>> Q. How Do I Reduce Switch Flooding Caused By NLB?
>>
>> "Distribution of Cluster Traffic" section in the article below:
>> Network Load Balancing Technical Overview
>> http://www.microsoft.com/technet/prodtechnol/windows2000serv/deploy/confeat/
>> nlbovw.mspx
>>
>> http://groups.google.co.jp/groups?hl=en&lr=&threadm=%23OYwFZZ%23EHA.3840%40t
>> k2msftngp13.phx.gbl&rnum=2&prev=/groups%3Fq%3Dunicast%2Band%2Bflooding%2Band
>> %2Bswitch%2BNLB%26hl%3Den%26lr%3D%26selm%3D%2523OYwFZZ%2523EHA.3840%2540tk2m
>> sftngp13.phx.gbl%26rnum%3D2
>>
>>
>> As for the "red question mark" on the network card when switching to the
>> multicast mode, where do you see the question mark, in NLB manager or in
>> Device Manager? Have you restarted the servers after the mode change?
>>
>> Please take a screen shot of the red mark and refer to the following
>> article gather MPSreport, please use MPSRPT_CLUSTER.EXE and
>> MPSRPT_NETWORK.EXE:
>> Microsoft Product Support's Reporting Tools
>> http://www.microsoft.com/downloads/details.aspx?FamilyID=cebf3c7c-7ca5-408f-
>> 88b7-f9c79b7306c0&displaylang=en
>>
>> Winzip all the info and send it to v-rebc@microsoft.com for research.
>>
>> Hope this helps!
>>
>> Best regards,
>>
>> Rebecca Chen
>>
>> MCSE2000 MCDBA CCNA
>>
>>
>> Microsoft Online Partner Support
>> Get Secure! - www.microsoft.com/security
>>
>> =====================================================
>>
>> When responding to posts, please "Reply to Group" via your newsreader so
>> that others may learn and benefit from your issue.
>>
>> =====================================================
>> This posting is provided "AS IS" with no warranties, and confers no
>> rights.
>>
>
>

Relevant Pages

  • Re: Windows 2003 Network Load Balancing Problem
    ... Sadly my servers websites were configured with specific IPs and host ... I find it very strange why the NLB driver can receive on a virtual IP ... he set his IIS website to use ALL UNASSIGNED addresses rather than ... specifically pointing it at the single virtual cluster address. ...
    (microsoft.public.windows.server.clustering)
  • Re: NLB Heartbeat/private network
    ... Configure Network Load Balancing (NLB) with Two Network Adapters ... Configure the NLB Cluster properties according to your parameters, ... Both machines have 2 NICs. ...
    (microsoft.public.windows.server.clustering)
  • RE: NLB heartbeat clarification please.
    ... Today I have an Oracle/Domino back-end cluster based on Win2k3 EE. ... question is how many public IP-addresses will I need (two servers - two NICs) ... > adapter is referred to as the Dedicated NIC. ...
    (microsoft.public.windows.server.clustering)
  • RE: NLB heartbeat clarification please.
    ... In describing your scenario with using NLB in the front-end and MSCS on the back-end, it sounds like you would need to follow the same ... Exchange Server 2003 and Exchange 2000 Server Front-End and Back-End Topology ... Windows NT/2000/2003 Cluster Technologies ... > question is how many public IP-addresses will I need (two servers - two NICs) ...
    (microsoft.public.windows.server.clustering)
  • Re: Load Balanced TS Servers and Unicast Flooding
    ... Remove each on of your tse servers from the cluster, ... > Use the other network card for cluster communication. ... >> NICs. ...
    (microsoft.public.windows.terminal_services)