Time for a techy post!
Came across this issue recently and thought i’d post it as someone may find it useful!
During build of some of the labs for EMCW we encountered an issue where the mgmt NICs on some of our nested ESXi5.5 VMs were becoming disconnected. It seemed to be confined to 5.5 hosts with E1000 NICs and only to hosts that had large amounts of resources (for nested ESXi hosts at least). These hosts had 32GB RAM & between 10 & 16 CPUs and have many nested VMs running on them. At random the ESXi VMs would lose mgmt network connectivity. During troubleshooting we first looked to see if they were being overloaded for resources and possibly suffering a panic, but when opening the console of the VM they were running just fine (aside from the mgmt network disconnected!!). Nothing obvious in the logs so we then started to look at the NICs. We have VMware tools installed thanks to this Fling so we decided to try some different NIC types. Being an old fashioned “one step at a time” kind of troubleshooter I started with E1000E with a view to moving to VMXnet3 however E1000E seems to have resolved the issue. No more disconnects!
Updating all of our base 5.5 templates with E1000E NICs & hopefully this is the last we’ll see of this issue!