Service operation is a collection of convenient Ops and management tools provided for components, including general service restart and command-based Ops operations specific to certain components, such as HDFS NameNode active/standby switch, HDFS data rebalance, YARN ResourceManager active/standby switch, and YARN queue refresh.
- YARN ResourceManager active/standby switch requires you to disable
yarn.resourcemanager.ha.automatic-failover.enabled. If it is not displayed in the operation drop-down list on the YARN block, you need to find
yarn-site.xmlconfiguration file of YARN configuration management and disable it.
- Do not delete the queues that have already taken effect in the scheduler's configuration file when refreshing YARN queues.
- Ranger metadatabase change is currently supported only for MySQL databases, and only the connections of admin users will be tested.
fair-scheduler.xmlto take effect in the ResourceManager. Do not delete queues defined in them.
conf/install.propertiesfile and then run the
setup.shscript locally. This script can sync the database information to the local
ranger-admin-site.xmlconfiguration file without modifying the content of
ranger-admin-site.xmlin Configuration Management. When you modify and deliver
ranger-admin-site.xmlbased on your needs, the database information will be overwritten, resulting in an exception. This operation allows you to quickly configure a Ranger metadatabase, so as to avoid exceptions caused by missing configuration items when you change the metadatabase address.