Cluster Dump Should Be More Resilient To Errors And Encode Errors Within Dump

[George-Aeillo]

Currently, within the cluster dump command, when the dump script is being run on a node, the set -e flag is set, causing the script to exit when it reaches an error. This is not ideal behavior, as the dump on the node to continue in order to gather more information. In addition to this, any errors that are found in a cluster dump, should be recorded in the dump itself, as that will allow users to better diagnose the state of their cluster and to find potential issues.