Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
New Member

NIC Redundancy Teaming/Bonding

Greetings all,

  We have a redundant TMS pair using TMS Version 13.0 and my question is more about adding another level of redundancy.

Our servers are running Windows 2008 R2.

Our Network Interface Card is a

Broadcom BCM5709C NetXtreeme II GigE (NDIS VBD Client) with 4 interfaces.

Is anyone out there using multiple interfaces on their TMS and SQL Servers to provide a level of redundancy in case the primary access switch goes down?

Currently we have the following:

switch A has TMS A and SQL Server connected to it

Switch B has TMS B connected to it (Currently we don't have a redundant SQL Server)

So if Switch A goes down our entire Video infrastructure is down because our SQL Server has not redundancy. 

We are in a production environment so I don't have the luxury of just trying it out.

I have asked the question of Cisco as to the ability to use multiple interfaces so that I can have TMS A have a second interface connected to Switch B

for redundancy and I have not gotten any feedback so coming to the community of users while I await an answer.

I am really looking to see if anyone is doing and if so, do you have a quick guide on how you set it up

Any assistance would be appreciated.

Thanks

Wally

3 REPLIES

NIC Redundancy Teaming/Bonding

Hi Walter!

Let me quote of the TMS Install and getting started guide:

Note:Cisco TMS cannot use multiple network cards on a server: and will only bind to the first available network interface. Cisco TMS can manage a public and private network so long as the public network’s connection is further upstream from Cisco TMS, rather than being directly connected to Cisco TMS (using multiple network interface cards).

So if you talk about two different ip interfaces I say: no.

If you talk about a network adapter which can have multiple connections to two switches and which

is transparent as one ip interface to windows, why not. I assume this is what you are talking about.

Wouldnt it make more sense to add the additional interface on the sql server and connect that one to switch B?

I have seen more servers die then switches, so what do you think of forching to have a redundant SQL setup as well?

If you have high demands on redundancy I would look into a virtualized platform with failover capabilities + nice redundant server hardware + redundant network infrastructure.

If you havent read it, take a look at the TMS Server Redundancy Config Guide.

I am not 100% sure which deployment model you try follow here, but in general for two TMS

which should be up at the same time you would need a load balancer upfront, or you have

some switch over time (like changing DNS records) anyhow, then you could also replug the

TMS in the other switch, ...

What is your critical application? Conference booking, Phonebooks, statistics?

Always consider as well how deep in detail you want to get and don't forget to secure

other parts as well (like the call control (vcs), MCUs, network access itself / router / ...)

Especially advanced setups should be tested, so you definitly should look into a lab setup, as well

as use a maintenance window one day to test if your failover really properly works.

Please remember to rate helpful responses and identify

New Member

NIC Redundancy Teaming/Bonding

Martin,

  Thanks for your reply.  I understand that we cannot have more than one Network Card in a server based on the documentation.  I come from a network background so I am not as server saavy as I would like to be.  Getting there slowly but surely.  Is each physical interface (4 of them) on the one NIC considered a seperate NIC?

Our TMS servers are using a load balancer as a front end so there is a level of redundancy there. 

I guess I could have explained myself better, or more specifically.  I am more looking to get some redundancy for the SQL Server setup.  We are currently working on getting a redudant SQL server as well, but that is a future endeavor so I am trying to work out solutions to provide a level of redundancy most specifically for that single point of failure we have in our SQL server. 

Trying to investigate the ability to add teaming/bonding to the servers.  This may be possible with the SQL Server but not the TMS Servers?  I have read about teaming/bonding, but not extensively, so I am looking to find out if folks are doing it and if so if they have some quick guides on setting it up.

Thanks again for your assistance by the way

Wally

Re: NIC Redundancy Teaming/Bonding

Hi Wally

The doc might better say one ipv4 and one ipv6 interface than "TMS cannot use multiple network cards on a server".

I would expect that the bonding part is transparent. So you still have only one IP interface and

the TMS application shall be fine with it.

How the ethernet layer and its failover is handled should not matter for the TMS.

So its more important what windows / NIC / switches can do.

The simplest might be a logical-ethernet bridge over  two ethernet ports of a server under windows

and connect it to two switches and use spanning tree.

But I assume there are better ways which often are vendor specific to your NIC/Switches.

So I would browse the docs and see what is supported.

Maybe your network / server colleges had implemented a similar setup for other servers.

Thx for the feedback. I would highly appreciate it if you could rate my messages by using the stars below each message!

Martin

Please remember to rate helpful responses and identify

792
Views
3
Helpful
3
Replies
CreatePlease to create content