Hi all,
I'm facing a very strange issue on CSG.
I have ten 7609 each with 2 CSG configured in fault tolerance. The alias IP address in client vlan of these CSG is used as next hop by a load balancer (7606 with SLB),the traffic is passing through the CSG to a set of SSGs that have default route in uplink vlan to a firewall load balacer that will route customers to many web servers.
It happened that suddenly all the CSGs in all chassis (simply load balanced,not talking each other) started to block only HTTP continuation back to customers.As far as I know, CSG rel4 goes mad when it receives tcp packet in wrong order since it has no buffer in server vlan,unfortunately, during the problem, all the sniff traces taken with ethereal showed correct TCP seq numbers coming back to CSGs from web server.This issue is related to a single IP address (defined with ip csg content) all other destinations are ok.The workaround is to restart all the CSGs.I already sent a show tech-support to TAC and I'm still waiting for answers.In the meanwhile, do someone knows how to understand why all the CSGs hanged simoultanesly?We got this problem on a PoP, after I week we faced the same problem on another PoP.The only common element is the destination IP of the web server.
Thanks in advance.
Regs,
Riccardo