Configuring GSLB for Active geo-site redundancy without using DNS

[jimmy.m.murphy]

Hi All,

I've been reading through the threads and the config examples on cisco.com looking to see if there is a way to have CSS11506 geo-site redundancy without using DNS?

Quick rundown of the issue:

1. Site 1, will have SMTP frontends, connecting to a geo-site redundant SAN.

2. Site 2, will have site 1's SAN replication software replicate to it's SAN.

3. Site redundant pair of CSS, at each geo-site.

4. Site's will have differing VLAN's most likely, but if required we can advise same VLAN.

When we have site 1 as primary, and all four frontend SMTP server's fail, is there a way for the CSS to redirect, or place VIP as down, so that backbone routers will auto redirect to site 2 CSS, which will act as secondary? Am unsure of backbone at this stage. Need to clarify this with client.

We're trying to provide active-active failover. It does not need to be stateful as the SMTP frontends will not be, but all data between SAN's will be replicated via separate path.

The aim of this question is to achieve automatic failover between sites, based on complete failure of frontends, but not necessarily failure of CSS.

Thanks in advance guys,

all help is greatly appreciated.

J

[jsirstin]

Jimmy,

You could create a redirect service and add this service to the content rule as a PrimarySorryServer. When all 4 SMTP servers fail the CSS would start handing out the redirect to the other location. When one or more of the SMTP servers come back online it will no longer hand out the redirect. Sample below.

Regards

Jim

service http-redir
type redirect
ip address 1.1.1.11
keepalive type none

no prepend
redirect-string "http://www.mydomain.com/"
active

content service-redirect-all
  port 80
  protocol tcp
  vip address 192.168.1.21
  url "/*"
  add service server1

  add service server2

  add service server3

  add service server4

  primarySorryServer http-redir

  active