Cluster restart also restarts Impala service. Could you please double check if following two properties are unchecked in CM > Impala > Configuration   Enable Audit Collection Enable Impala Audit Event Generation To be able to disable auditing for Impala service, you need to uncheck both of the above options. Once done make sure to restart Impala service. CM would show stale and restart icon as well.   ... View more

Hi @pawski,    Here is a solution for comparing CDH clusters configurations managed by different Cloudera Managers  https://community.cloudera.com/t5/Support-Questions/Compare-settings-in-different-clusters/m-p/280536/highlight/true#M208909    Thanks, Salim ... View more

Hi Devendra,   I see you are seeing multiple low disk space alerts for various role log and role data directories. It is not advised to delete anything manually from these directories instead adjust the retention policy for them in CM and let CM delete it automatically for you.   Review the documentation to find default retention policy for Cloudera Management Services data directories and following documentation for setting retention policy for any role log that by default reside in /var/log/. Remove and Delete any unused parcel from CM > Parcel page to reclaim disk space occupied by parcels. Details are given here.    I hope that will help.   Thanks, Salim  ... View more

Hi All, Unfortunately, other than CM API there is no feature in CM itself to compare the cluster configuration with other cluster managed by another CM instance. In another case if you want compare the configuration settings for a particular service between two different clusters managed by same CM instance, you can do as per the steps given in the documentation here [1]. I suggest to go with CM CPI solution. You can use the CM API [2] to retrieve the full cluster configuration in JSON format including all hosts, clusters, services, roles, users, settings, etc. and use third party text compare tools like this [3] to compare them as text. It will highlight the difference in yellow. Another way is using diff command line.   e.g. CM API example for getting full cluster configuration in browser in JSON format   http://<your_cm_hostname>:7180/api/v12/cm/deployment   Get the full cluster configuration using the API from both the CM servers and put them into the text comparator.   I hope this will help.   Regards, Salim [1] https://docs.cloudera.com/documentation/enterprise/5-13-x/topics/admin_compare_service_config_across_clusters.html#xd_583c10bfdbd326ba-7dae4aa6-147c30d0933--7c7c [2] https://cloudera.github.io/cm_api/apidocs/v12/path__cm_deployment.html [3] http://iblogbox.com/devtools/diff/ ... View more

Hi, The error message confirms that the LevelDB holding the YARN state store is corrupt : org.apache.hadoop.service.ServiceStateException: org.fusesource.leveldbjni.internal.NativeDB$DBException: Corruption: 1 missing files; e.g.: /var/lib/hadoop-yarn/yarn-nm-recovery/yarn-nm-state/000042.sst Solution is to clean up and recreate the state store database : 1. In CM make sure the affected NodeManager has status STOP 2. Backup the contents under /var/lib/hadoop-yarn/yarn-nm-recovery/yarn-nm-state to a different directory. 3. Delete all the contents under /var/lib/hadoop-yarn/yarn-nm-recovery/yarn-nm-state. 4. Start affected NodeManager. ... View more