Restarting a dead service on a node
hbase regionserver logs were saying they were timing out talking to hdfs ports. This led us to a problem with the “datanode” service, and that above blog about restarting services is how we restarted that service.
Just looking through these will help to find what’s broken. var/logs tmp