The whole process of HDP upgrade can be divided into two steps:
· setting up the http server so that we can access the files across the network
· setting up local repos on all the nodes to install the HDP rpms with dependencies
Setting up python http server
With the help of the python http server all the nodes are able to access the HDP rpms. You can set up the http server with the help of the below commands:
· Python 2 — python -m SimpleHTTPServer 8000
· Python 3 — python -m http.server 8000
you can go to the /var/www/html directory and create a folder with all the HDP and HDP utils rpms.
Once we set up a python http server, we can access the folder created in the location /var/www/html from any machine in the same network.
Verifying that the fileserver is running:
· netstat -anlp |grep 8000 — since we are hosting the files which would be available it the server we set up the python http server and the port which we used while setting it up, in our case it is 8000
· we can ps -ef |grep processid — the process id running on port 8000 we get from the netstat executed above
· We need to pass the package paths and register the new HDP version in ambari and save it
Setting up the local repos
we set up the local repos in all the nodes so that HDP and HDP utils rps are installed with all dependencies on all the nodes
Create a repo file in the /etc/yum.repos.d
the repo file should have the below info:
Repository ID, Name of the repo,base url,enabled and GPG check
GPGcheck — we can enable or disable the gpgcheck
The format the repo file is as below:
once the above two steps are performed, we initiate the upgrade from ambari. The HDP packages get installed on all nodes.