I think you should be able to implement this kind of logic using default LTM monitors with the pool config set for a minimum number of successful monitors. You can hard code the destination port on each monitor and then associate all four of the monitors on the pool.
Not sure if I can do this using default monitors with the Pool config. On the Pool I can select all or at least for the number of monitors which need to be up. What I need to do is say the http monitor always needs to be up and only one of 4 tcp port needs to be available.
So if said I need at least 1 monitor need to be available I could have a situation where the http monitor is down and one tcp port is responding which would mark the Pool member as available when infact it shouldn't be.
Sorry if I'm missing something here. You might remember me from a training session you did in North Yorkshire last year.
On another note is it easy to create an i-rule which states always use this pool member x unless it goes down then use otherwise pool member y? We've created a pool two load balance outbound internet connectivity via our 2 proxies but I have an app which needs to use the same proxy unless there is a failover. My thoughts are I could create a dedicated VS for this and apply an i-rule to it. To get round this I've had to create two pools which I'd rather not do due to the external monitors created to monitor the internet links.
Nice to hear from you. I don't see any of your private info on DC, so I rarely know who I'm replying to :)
I wonder if you could hard code the destination port for the HTTP monitor and associate that with the node address. Then add the four TCP monitors to the pool and set a min success of one. LTM should then mark the node address (and all pool members defined on that IP) down if the HTTP monitor fails. If all of the TCP monitors fail, then the pool member would also be marked down. Does that sound like a plausible solution?
I'll reply to your other post on proxy monitoring as you have more info in that one.
Right understand now, yes I think that will work. My concern was the nodes are members of two different Pools but having checked they are using the same http monitor on each Pool. Thanks very much for your help.