b)In case of “clssnmPollingThread” after the node is evicted does it reboot?

c) assuming above error is for network heartbeat what doe the error look like when it is vote disk io timeout.

Thanks

Oracle Clusterware doesn’t have an ability to support redundant interconnect natively so the way to configure redundant interconnect is to use a third party product. There are quite a few terms for it like bonding, trunking and so on.
In a nutshell, all those products join more than one network cards and represent it as a single virtual network card so that if one of real network cards/cables are not operating, it’s transparent for virtual network card (except lower bandwidth). This way Oracle Clusterware is just using one network which is redundant behind the scenes.

What platform you are on? What technology you are using?

To test if it works I would start with setting a simple ping over interconnect from every host. Before doing so — stop clusterware on every node using “crsctl stop crs”. Then while pings are running — pull out a cable and see what happens with ping responses. Wait several minutes to see if it recovers – perhaps, the technology you are using takes long to reconfigure.