Hi, I have a cluster witn CDH 5.4.2, and the whole cluster is working fine, but when I try to access to the Hue Web UI doesn't work, ( https://host:7180/cmf/services/23/status ) when I go to https://host:8888 , it doesn't work.. ( it looks like a ERR_CONNECTION_TIMED_OUT error ) I got the next error at the Log from Hue.
5:37:45 AM INFO settings Welcome to Hue 3.7.0
5:37:46 AM WARNING __init__ Couldn't import snappy. Support for snappy compression disabled.[09/Jun/2015 20:00:13 ] settings INFO Welcome to Hue 3.7.0
6:00:14 AM WARNING __init__ Couldn't import snappy. Support for snappy compression disabled.
6:00:14 AM INFO settings Welcome to Hue 3.7.0
6:00:15 AM WARNING __init__ Couldn't import snappy. Support for snappy compression disabled.
do you have any idea ?
In putty connected to the master instace and performed ping gke-hadoop-ubuntu-default-pool-78514257-626p.c.platform-services.internal ..... its responding. Also performed curl http://gke-hadoop-ubuntu-default-pool-78514257-626p.c.platform-services.internal:10002/ (my hue server is down so tried connecting to hive having 10002 port) got this:
I think in putty its connected but in browser its not resolving the hostname....
I'm talking about DNS.
For instance, if using a DNS server for name resolution:
$ nslookup www.google.com
My point in this case is that someone got an error saying that the browser could not resolve the IP.
For example, see:
The browser problem showed an issue with the browser and how it tries to communicate with the server host; it had nothing to do with Hue.
Actually, that stackoverflow may be off target since it talks about public addresses. Your browser needs to resolve a hostname into an IP so that your computer can connect to Hue. If that resolution (or mapping) fails then the client (browser in this case) cannot connect to the host specified in the browser's URL.
I got these WARN in kt renewer under hue log and after two hours of troubleshooting learnt that the PID on Cloudera Monitoring serverices had some conflicts.
Had to stop Cloudera Monitor services
restart cloudera-scm-server and agent
It did all the magic.
It might be different for others, though.