critical reporter and critical service

Unanswered Question
Jun 20th, 2007

Good Day CSS Experts,

I am configuring CSS 11503 in VIP and Virtual interface redundancy.

Client----L3S/w--- CSS1----CSS2 -----L2 (stack)--server1

I have understood that rather using upper and lower device ip for monitoring, we should use CSS interface IP in the critical service for best & fast failover result.. So I plan to use critical service to monitor CSS Int. & define the CSS interface IP addresses in the keepalive type script ap-kal-pinglist "client_side_CSS_int_ip Srvr_side_CSS_int_ip"

1. What IP address should I assign to the service used for critical server? I can define the above keepalive but, the service upstream_downstream_CSS, does not accepts the CSS IP address and so I can?t activate the service without IP address?

******************

I understand that we can not use critical reporter vrid-peering & physical interface at the same time.

2. Should I use a critical reporter using vrid-peering or physical interface with the critical service I mentioned above?

Goal is to have best failover if any bit goes down with the related environment L3----CSS---L2---server1

Also, any best working solution opinion welcome!

Thanks,

Gagan

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
gagansethi Sun, 06/24/2007 - 17:24

Good Day,

Any suggestions on the above shall be appreciated.

Also, do i need to have critical service and critical reporter configured on the standby CSS as well ?

Thanks,

Gagan

skumar1969 Tue, 06/26/2007 - 20:25

Sethi,

In your case, what you are trying to achieve is a topology called "fate sharing". A CSS has 2 circuits, a client side and server side. When we use 2 CSS as a failover pair, there were situations that client circuit of 1st CSS coupled to server circuit of 2nd CSS resulting in an asymmetric traffic flow lead to traffic disruptions. This is no good and we do not want this to happen. To avoid this we go for a solution using critical services and reporter services. You know what is a reporter service as I can see you had used it in the right sense.

When you use 'critical services' you need to monitor ip addresses that is always outside the CSS box. We then would use 2 ip addresses, one for each circuit.

1/ On your 'client side' circuit you need to monitor an ip usually your CSS's default gateway.

2/ On your 'server side' circuit you need to monitor an ip (usually the management ip) of the L2 switch.

Example:

service upstream_downstream

ip address 10.30.31.5

ip address 10.30.41.15

active

Yes on both the boxes configure the cri services same way, same ip addresses etc.

However on the reporter config will have different ips as per the circuit interface ip.

Hope this would help. Good luck!

thanks

Actions

This Discussion