cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1296
Views
15
Helpful
19
Replies

Documentation on Catalyst 6500 Chassis Failover types

new_networker
Level 1
Level 1

Could you please provide me detailed documentation on Cisco Catalyst 6500 Chassis failover. It may cover

1) Failover of complete Chassis (along with all service modules)

2)Failover of independent service modules across different Chassis (i.e. few modules may be active on Chassis A and remaining modules may be active on Chassis B).

Thanks.

1 Accepted Solution

Accepted Solutions

It is for all vlans that have end hosts in them so yes you would need to configure all SVI's as such.

Jon

View solution in original post

19 Replies 19

Jon Marshall
Hall of Fame
Hall of Fame

You do like these big questions don't you :-)

http://www.cisco.com/en/US/products/hw/switches/ps708/index.html

The above is a link to the 6500 switch page. On here you can find an awful lot of information with links to white papers on architecture/high availability etc , configuration and command references, Q&A's and a lot more.

There is so much that could be covered on module failover, service module (FWSM/ACE/CSM etc..) failover, supervisor failover both within the same chassis and between chassis and then full chassis failover that it's hard to give a succint answer here.

Have a look at some of the links and please feel free to come back with as many queries as you like. There are some very experienced 6500 engineers in this forum so you should always be able to get an answer.

If there was anything in particular you were looking for info on let us know and we may be able to point you in the right direction.

Jon

The given link has plethora of information covering almost everything. Could you provide me the link to a single document (if any) with failover types. I haven't been able to find one.

Thanks.

You haven't been able to find one because the information is scattered throughout many docs, that's part of the problem.

http://www.cisco.com/en/US/prod/collateral/switches/ps5718/ps708/prod_white_paper09186a0080088874.html

The above is a link to supervisor failover both within and between chassis. It covers the options you have in terms of high availability.

For the service modules the explanation of failover will be found within the particular service module documentation.

As far as ethernet modules go if the module fails then it entirely depends on what is connected into it ie. if you have a server connnected to an ethernet module in one chassis and the same server has a connection to another ethernet module in a 2nd chassis and these 2 chassis are interlinked at L2 then if the module with the active server NIC connection fails the server will just switch to it's other NIC.

I don't want to overload you with information but you may also find the SRND (Solution Reference Network Design ) docs of some use for this.

www.cisco.com/go/srnd

Have a look at the Data Centre section which has papers which cover common failure scenarios with 6500 switches.

Jon

The provided link doesn't mention about inter-chassis (dual chassis) failover procedure.

In my case, there are two 6500 chassis each holding single SUP unit. Do you have a link for material related to such scenario.

Rgds.

Try starting here: http://www.cisco.com/en/US/prod/collateral/switches/ps5718/ps9336/prod_white_paper0900aecd806ee2ed.html. You can then search Cisco's site for further information on VSS.

Anything specific for WS-SUP720-3B, as the VSS is only applicable to VS-S720-10G-3G/XL.

You need the later VS-S720-10G-3C(XL). (You might investigate whether there's any trade up option, available to you.)

So in other words, are you saying that inter-chassis (dual chassis) failover is not possible for WS-SUP720-3B.

VSS is a relatively new technology that in effect allows the 2 chassis to be seen as one. However this is not quite the same as talking about inter-chassis failover.

If you are asking what happens when one of the 2 chassis fails then as i said it all depends on what you have running in the chassis in the first place. Lets take a relatively simple example

2 6500 switches connected by a layer 2 trunk with 10 vlans.

Servers are dual connected to both switches.

6500_1 is the spanning-tree root and HSRP active switch for all 10 vlans.

A WAN router is dual connected to both 6500 switches with L3 Point-to-Point links. Traffic from users comes from the WAN via the router to the servers.

So lets say a user requests an http page. The request comes to the WAN router, the WAN router sends the traffic to 6500_1 which then forwards the traffic to the http server.

The http server then sends the traffic back to 6500_1 because that is where it's active HSRP gateway is, 6500_1 sends the traffic back to the WAN router which sends the traffic back to the user.

6500_1 chassis fails. The servers are all dual-honed so they now start to use their other NIC's which are connected to 6500_2.

HSRP moves the active gateway across to 6500_2 and 6500_2 becomes the STP root for all 10 vlans.

Now another request comes in, the WAN router sends the traffic to 6500_2, 6500_2 sends the traffic to the http server. The http server then sends response back to it's active gateway which is on 6500_2, 6500_2 sends back to WAN router and WAN router to user.

Note - if you have 2 L3 links from the WAN router to 6500_1 and 6500_2 the traffic path i described may be slightly different in that the WAN router will see 2 equal cost paths to the server vlan (before the chassis failure) and so may send to either 6500. If it sends to 6500_2 then 6500_2 simple switches the traffic across the L2 trunk connecting 6500_1 & 6500_2. The key point i was trying to get across was how the servers react to a chassis failure.

Jon

Voila! That was great.

Again, is there any documentation on HSRP and trunking configuration across two 6500 chassis.

Thanks.

Is NSF/SSO a better option than HSRP on Cat 6500 ?

Two different things.

NSF/SSO is related to Supervisor failover and how quickly it can achieve this failover. So this will affect how quickly HSRP reacts. But NSF/SSO cannot be used directly to provide server redundancy.

HSRP is concerned with providing a redundant gateway to end hosts. So on

6500_1

int vlan 10

ip address 192.168.5.2 255.255.255.0

standby 10 ip 192.168.5.1

6500_2

int vlan 10

ip address 192.168.5.3 255.255.255.0

standby 10 ip 192.168.5.1

From previous thread for the server to have redundancy it would have it's default-gateway set to 192.168.5.1. So if 6500_1 fails and it switches to 6500_2 the server can still use 192.168.5.1 as the gateway address.

The alternative to HSRP is GLBP which allows both switches to be active gateways at the same time.

In answer to your previous question, if you just want to know how to configure HSRP, Layer 2 trunks, etherchannel links start with the configuration guides -

http://www.cisco.com/en/US/products/hw/switches/ps708/products_installation_and_configuration_guides_list.html

Also some of the design guides on Data Centre Infrastructure and Campus Design have sections concerned with redundancy in terms of HSRP, routing protocol convergence (which we haven't really touched on). At home i have broadband but i'm afraid i am on dial up at the moment otherwise i would do a more specific search among the design docs for you.

Jon

"Is NSF/SSO a better option than HSRP on Cat 6500 ?"

"Better" is very subjective. Jon touches upon the speed of fail-over, but another fail-over issue is often whether state infomation is maintained, also whether other devices perceive one device or multiple devices. These issues can be especially important when you dealing with various services modules.

I didn't intend to imply VSS is the only method to deal with failure of a 6500 chassis in a dual configuration (and may have done so - sorry), but it has advantages somewhat like dual sups within a single chassis. Its biggest disadvantage is its many limitations since it's so new, such as current lack of support for many service modules.

Jon, HSRP is not supported on SUP720. I believe NSF/SSO is the answer.

Regards.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco