The same issue happened to me. I was able start hue successfully after increasing the number of CPUs in VirtualBox.
I also increased the amount of RAM earlier. The original CPU I had was 1, changed to 3
Increase the memory of docker to 8GB if you can. Otherwise, set it at least 4GB.
Let hue fail while starting the container.
After that, attach to the docker container and access its shell to run the following command,
I have encountered the same issue here, and resolved now based on the comments and posts above. There are two issues mentioned above:
Failed to start Hue.
In my case, this is caused by limited resources allocated with default docker VM settings. According to @Ronald Teo's answer, going to
VirtualBox -> 'default'[your docker-machine name] -> Settings ->
, increase base memory to 8192MB, and processors to at least 3, have fixed my problem.
Can not access Hue from my host machine. Based on the original post, Try docker run --privileged=true --hostname=quickstart.cloudera -p 7180:7180 -p 8888:8888 -t -i 9f3ab06c7554 /usr/bin/docker-quickstart should solve this problem.
When you run docker using -p 7180 and -p 8888, it will allocate a random port on your windows host. However, if you use -p 7180:7180 and -p 8888:8888, assuming those ports are free on the host, it will map them directly.
Otherwise you can execute docker ps and it will show you which ports it mapped the 7180 and 8888 to. Then in your host browser you can enter
If its all on your local machine, you shouldn't need the port forwarding.
Since you're running the docker machine inside a VM, you need to open the port on VirtualBox.
You can do this from the Port Forwarding button in the network adapter panel in VirtualBox.
Settings > Network > Advanced > Port Forwarding
You should see an SSH port already being forwarded for docker. Just add any additional ports like that one.
And here are lists of all the ports used by CDH. Of course you don't need all of them. I would suggest at least Cloudera Manager (7180), namenode and datanode UI (50070 & 50075), and the job servers like mapreduce (8088,8042 & 10020) or spark (18080 & 18081). And I personally don't use it, but Hue is 8888.