Health Monitor Depedancies

Hi All

 

 

I hope my question is advanced enough to ask in this forum. if not i apologise. I have manage a pair of LTM's running 10.2 which load balance traffic to a tiered application architecture, which basically flows like this

 

 

web Traffic --> web VIP --> Web servers --> application Vips (1 vip for each service at application layer) --> application servers --> database vip --> database servers

 

 

for a long time the environment was using the most basic health monitors (icmp, or a tcp half open heath check) for each application pool assiocated with the vip at each layer and no dependancy on the layer below. so if something serious took out the application tier to, the web vip would continue to process requests and they would fail because the application wasn't there to process the requests it was receiving from the web servers.

 

 

I have recently implemented a GTM which allows me to fail over 2 a secondary site in the event of the web server vip on the ltm being down. what I would like to have occur is in the case of a failure at the application or database tiers the web vip be marked as down which would detected by the gtm and would then redirect traffic to our secondary site.

 

 

I need some advice on the best way to build some dependacy checks on the lower tiers, I have some rather restrictive limitations as to what I can publish to the gtm in terms of ltm - gtm sync using iquery so i was wondering if it's possible to build this dependancy model on the ltms

 

 

I hope this has made sense, and I would appricate any advice you can offer

 

 

Many thanks

 

Andrew