Ran into a new issue today. Was trying to make some wiring changes to our network cards on our VSphere servers. We have two VSphere 4.1 servers, each have 6 NICs, with 3 virtual Switches, with 2 NICs per virtual switch for redundancy, each part of a single HA Cluster in vCenter. One virtual switch is for our iSCSI traffic, one is for all of our VM traffic, and the final virtual switch is just for VMotion and the Service Console. My plan was to start with the service console and unplug one NIC, run a new wire, and plug that NIC back in. Then once that NIC showed as operational in the VSphere Client, I would unplug the other NIC, run the new wire, and plug it back in.
I started on the first server and we unplugged one NIC from the service console vswitch and ran the new wire and plugged it back in. I saw the NIC become active in the VSphere client again so we moved on to the next NIC. After unplugging that NIC and running a new wire and plugging it back in, i noticed that all the VMs were shutdown. Not good! So I powered them all back on and brought everything back online.
I then decided that i shouldn't run the new cables for the other server till off hours but i was confused as to why all the VMs shutdown. So i started to do some troubleshooting on the next server by just unplugging one of the service console NICs and waiting 1-2 mins till the VSphere client reported the NIC as down. Then once i saw it was down, i plugged it back in. As soon as i plugged it back in, all of the VMs on that server proceeded to shut down!
Can anyone enlighten me what setting in my HA Cluster in VCenter i have wrong, that causes all my VMs to automatically shutdown when i lose a NIC to my service console vswitch?