I was trying to play around with Hadoop ecosystem. One way to do that is use the Sandbox-HDP from Hortonworks (merged by Cloudera now).
Here s the document for the installation with Docker. link
As easy as this may looked, I still got two problems while working on it.
Here is the solution to the problem I had.
sandbox-proxy can’t be generated successfully.
It fails when running sandbox/proxy/proxy-deploy.sh with the error message as following.
docker: invalid reference format: repository name must be lowercase.
It’s a dumb problem actually. I found out that, since proxy-deploy.sh is auto-generated with user’s path, and my user name contains space. There is no escape in the path that generated, so simply add a escape before the space for every path in proxy-deploy.sh file.
sandbox-proxy crashes after started
After fixing the first issue, both sandbox-hdp and sandbox-proxy containers are created successfully. But sandbox-proxy always crash soon after started with log shows[emerg] 1#1: host not found in upstream “sandbox-hdp” in /etc/nginx/conf.d/http-hdp.conf:9.
It turns out that although the install guide says run this command sh docker-deploy-{HDPversion}.sh . The part of the script below requires single “=” for string comparison, otherwise it should be run with bash, meaning bash docker-deploy-{HDPversion}.sh is what we should use to run the deploy script if keep the script unchanged.
if [ “$flavor” == “hdf” ]; then
hostname=”sandbox-hdf.hortonworks.com”
elif [ “$flavor” == “hdp” ]; then
hostname=”sandbox-hdp.hortonworks.com”
fiIf everything works fine, then now sandbox-hdp and sandbox-proxy containers should be running, then simply type http://127.0.0.1:1080 in web browser. Now we will see the expected page. Finally.
Now we can use maria_dev/maria_dev to login and have fun!
Here is another video for the installation guide. link
It shows how easy it could be if none of the issue happened. lol
