On 30 January 2013 16:56, Michael Hanselmann <[email protected]> wrote: > 2013/1/30 Bernardo Dal Seno <[email protected]>: >> node = qa_config.AcquireNode() >> try: >> + old_es = qa_cluster.TestSetExclStorCluster(False) >> qa_cluster.TestExclStorSingleNode(node) >> >> - old_es = qa_cluster.TestSetExclStorCluster(True) >> + qa_cluster.TestSetExclStorCluster(True) > > You should use try/finally to handle errors.
Actually we don't clean up the cluster state after an error, and I don't think this should be an exception. In this way you can easily log-in to cluster and see what's the state and try to re-run the commands yourself. Bernardo
