Issues adding a single node cluster into management by VMM

Recently we have seen an issue where a few customers have been experiencing errors and long delays in adding a single-node (i.e. a 1 node cluster) cluster under management in Virtual Machine Manager. The error that is displayed in the console is Warning (13926) that all of the nodes could not be contacted.

This issue stems from the fact that VMM has an overcommitted property for each cluster. By default, a single-node cluster will always be overcommitted since any failure in that single node will cause the entire cluster and its Virtual Machines to go down. Such a failure does not go well with the concept of high availability.

There really are two workarounds for this issue:

1. Set the node reserve failure to 0 in the cluster properties. Unfortunately this can only be modified once the cluster has been brought under management by VMM. 

2. Add a second temporary node to the cluster simply for the purposes of adding it to VMM. This second node does not need to be physical hardware, it can be a VM. Then once the cluster is under management the second node can be removed.

 

If the single node cluster is added normally, it will eventually be added to the VMM console. However, this process can take a very long time. To expedite this process, it may be necessary to cycle the vmmservice.

Comments

• Anonymous
  July 31, 2013
  Thanks.... now able to continue testing my one-node cluster in the lab.