GSS TCP Keepalives

[jmiller]

I have a question concerning the overall design of the GSS product line as I am still a bit new to the design of enterprise level DNS loadbalancing. We have two GSS 4490 version 1.2. Recently we have exceeded the TCP keepalives for fast detection. The upgrade path to 1.3 still will not address the keepalive issue. My question is what is the best design to address this issue while keeping the fast detection? Or is it not a option at this point to keep fast and just use the standard detection method and values?

My thoughts were to setup a new pair and split the sites between the two. But this opens the can of worms for the daily operation of these systems, is there a central management interface that we could utilize?

[#2]

To designate the current interface as the interface that will be used for GSS keepalive communication, use the gss-tcp-keepalives command.

gss-tcp-keepalives

The first Ethernet interface (eth0) is used for keepalive traffic by default.

The gss-tcp-keepalives command is part of the suite of interface commands for the GSS and can only be used along with the interface ethernet command.

You cannot enter interface commands while the GSS is running (for example, serving Domain Name System [DNS] requests). You must enter the gss stop command before entering the gss-tcp-keepalives command