There are several options for you here, you have to pick one that best matches your requirements. If you are looking to backup your HDFS data but don't have capacity for a DR cluster, then backup to AWS S3 may be an option for you. Requires a cluster upgrade to later 5.x version, though, as this option is not there on 5.4 version.   Besides that, if you are able to completely restore the cluster nodes from your filesystem backups then you will be able to restore your HDFS data.  ... View more

Starting/stopping cluster or services (and many other actions) will be done in Cloudera Manager UI or using the Cloudera Manager API    The CM API can be used from curl commands in a shell script or when using the bindings from a Python script or Java code. ... View more

Try something like  # curl -u admin "http://$cmhost:7180/api/v33/clusters/Cluster%201/services/IMPALA-1/impalaQueries?filter=(query_state=RUNNING%20and%20query_duration>=1000)&from=2021-05-04T00:00:00.00Z"   ... View more

Thanks @SuryaDilip    Lets try for a test to set CM -> Administration -> Settings -> Category: Ports and Addresses to the actual FQDN of the CM host. This needs a restart of CM server.    After that change, update Alert Publisher config and re-enter the mail server hostname again, and restart Alert Publisher. Do you see the same exception in the Alert Publisher logs then, for a test alert? ... View more

Thank you. That confirms alert email delivery fails because there is no SMTP daemon listening on localhost:25.    I suspect there is a name resolution issue. Is the mail server hostname the host where CM is installed on, or a different host? Please login to CM, change the url in browser location bar to /cmf/debug/jmx and grep for MBeanServerId, does that show localhost or the correct hostname?   What do those commands show: # getent hosts <mail server hostname> # getent hosts <cm server hostname> # getent hosts localhost   ... View more