create 'check cluster recipe' Makefile trigger, ideally more detailed and thorough, almost a provision no-op where possible

[ssenator]

if regular provisioning doesn't end cleanly, per node or per cluster, suggest this be triggered in a message to the user

[nfrumkin]

some suggestions for functionality:

• it may be good to incorporate a check on files in the useradd/verify/vc* directories to make sure that the users have properly added a driver script/whatnot.
• verify that .out files are not interfering with the normal provisioning process. There may be straggler .out files in certain directories
• make sure cluster nodes are in the correct state (ie make sure that vagrant global-status agrees with the contents of ._state/provisioned)