I have a tcp farm up and working with 2 backend's. I created a custom 
farmguardian script that I can toggle to put servers in / out of the loop. If I 
failover the backend's (swap 1 in the loop and the other out) and start new 
connections they go to the in the loop backend as expected. One interesting bug 
I found is that if I create 200 socket connections in sequence after this 
failover (mode is priority / weight and both servers weight / priority 0 but 
only 1 is ever active at the time) I see most going to the new back end however 
some at random connects will be sent to the backend out of the loop. This 
symptom can be fixed by checking persistence it seems...

The problem I'm having that I haven't found a solution for is in regards to 
active socket connections during a failover. If I swap the backend's (failover, 
keeping both backend's running but the farmguardian script takes one out by 
showing it's down) all new connections go to the new backend as they are 
supposed to but the active sockets are maintained. This means that the software 
never gets a socket closed exception (trigger to reopen one which would go to 
the new backend) nor does the traffic get re routed to the new backend.

How can I have it so the active connections get killed or rerouted without 
killing the actual backend server (the one out of the loop)?

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support

Reply via email to