My client is using CDH 4.3. All services (hdfs,mapreduce, hue, hive, oozie, zookeeper) are in good health, except the hue service, which states hue server running and beeswax not running - it's reporting down. I can successfully connect to the Hue UI and run hive queries via the Hue UI's Beeswax i/f. So I know beeswax is not down. On CLI, I bring down hue and it still reports hue server and beeswax running; seems the hue and beeswax processes were spawned by supervisord (/usr/lib/cmf/agent/src/cmf/../../build/env/bin/python /usr/lib/cmf/agent
Oops I hit "me to" accidently.
Given you actually restarted from the command line, it could be there is a mix of installed CDH deployment types. CDH has 2 modes of deployment, Parcel and Package. If parces are deployed but packages were present first and not removed, the package based installation can start first on system reboot, preventing the parcel based service from starting.
Verify if parcels are being used in the CM UI, and if so then check if you have residual packages present that are starting as a service.
rpm -qa | grep cdh
rpm -qa | grep hue
rpm -qa | grep hadoop
If they do have a mix of pacakages and parcels, the CM 4 installation guide covers the proper removal of the pacages (rpm or debian).
CDH Hue integration on 4.3 is minimal as well. At some point they will want to evaluate catching up with the current release distribution of CDH either on the 4.x or 5.x releases.