I have attempted to increase the local storage of a VSA node as per the ESXi 5.5 / VSA 5.5 documentation. I chose the "full replacement" method which involves removing the node from the cluster, deleting the VSA appliance, adding hdd/increase physical storage and reinstall ESXi on the new, increased raid array. The next step is the reintroduce the node to the VSA cluster my selecting "Replace Appliance " in the VSA management tab. This is where I ran into problems.
The HP servers I'm using have a single HP quad port NIC installed. VSA normally requires a minimum of 2 NIC's the pass the install tests however there is a workaround for using a single quad port NIC which involves configuring the VSA networking on the nodes before deploying the VSA install. This workaround was successfully employed to install and create the VSA cluster intially. After reinstalling ESXi on the replacement node I have again setup the VSA networking as required and attempted to reintroduce the node into the VSA by selecting "Replace Applaince" in the VSA management tab as per the documentation. The replacement wizard launches and the process begins but it failed giving an error, I can't remember the exact error but after investigation it seems to relate to failing the networking test. The wizard said it would roll back any changes, after refreshing the VSA management tab it seemed I could again select "Replace Appliance" and did so. This time the wizard hung for a quite a while and again failed. Now the entire VSA cluster is showing offline (although I can still access the datastore on the remaining node ) and I can't select "Replace Appliance" anymore, it's greyed out. It's possible I configured the networking incorrectly and would like to reattempt the replacement but can't in the current state.
Any suggestions on how I can get the VSA cluster back to a state that will allow me to reintroduce another node again? I have moved all important VM's off the VSA storage and if necessary I could replace the whole cluster but I would prefer to be able to reattempt a reintroduction of the replacement node.