I believe in HDP sandbox Ranger audits are going to Mysql (not Solr). To change audits to go to Solr you would need to; 1. Start Solr and create the ranger_audits collection. The steps for this would differ whether you want Solr to run in standalone or cloud mode Standalone mode: https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.4.0/bk_Ranger_Install_Guide/content/solr_ranger_configure_standalone.html Cloud mode: https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.4.0/bk_Ranger_Install_Guide/content/solr_ranger_configure_solrcloud.html 2. change the ranger.audit.source.type to solr (from db) under Ambari > Ranger > Configs 3. Set appropriate property to point Ranger to Solr e.g. in Solr standalone mode, set ranger.audit.solr.urls = http://(your solr host fqdn):6083/solr/ranger_audits in Cloud mode, instead of providing url you would set ranger.audit.solr.zookeepers = Zk list e.g. ip-172-30-0-180.us-west-2.compute.internal:2181,ip-172-30-0-182.us-west-2.compute.internal:2181,ip-172-30-0-181.us-west-2.compute.internal:2181/ranger_audits 4. Restart Ranger service via Ambari ... View more

Does ranger creates unix groups during AD/LDAP sync? No - the usersync just brings in the users and groups for you to see and to be able to create Ranger policy based on the known users and groups . It does not create them it just reads from your defined source be it unix , AD/LDAP . Curious if the unix groups are used (based on sync) for authorization or native AD/LDAP groups. You create policy and this will let you control access not authorization. The underlying linux filesystem still needs to have SSSD or winBind/samba setup to show the same groups on the filesystem and the group names need to be the same . Ranger User sync will not create these groups in linux or hdfs. ... View more

When a batch of write-ahead log (WAL) edits are sent from the source cluster to the destination (active-active or otherwise), a response is sent to the source cluster from the destination. If the replication was successful, the response is essentially empty. Any errors encountered during the replay of those WAL edits would be returned to the source cluster. ... View more