Hello Geetha, Thanks for work done in this charm. We appreciate your efforts in to bring WebSphere to the Juju community! There are many things are done very well in this charm. It passes `charm proof` without warnings or informational messages! # README.md This readme contains multiple spelling errors and markdown formatting problems. Please go to the charm store to see how the markdown renders to HTML https://jujucharms.com/u/ibmcharmers/was-base/trusty This is the “ibmcharmers” personal namespace that is updated when you push code to launchpad. Use this page to see if the markdown renders correctly. I was unable to follow the readme to successfully deploy this charm on amazon because the virtual machine ran out of disk space. The Juju creates default amazon instances of 1.7GB of RAM and 8GB of root-disk. An installed instance takes over 15GB of root-disk! The readme should include some information about how to use “constraints” to request a larger root disk. Also you might want to add something about the suggested RAM memory to run WebSphere correctly. The “Creating the repositories in Apache” section could contain more detail and perhaps some examples of a url and what the file structure looks like when the apache is set up correctly. Please consider someone new trying to read these instructions. Can they follow along and set up a correct repository? Would they be able to deploy? The steps to deploy the charm are missing the `sha_im` configuration option. Again examples of valid commands would be useful here, I do realize that IP addresses change, but some of these values can be given as example: juju set was-base accept-ibm-im-license=True accept-ibm-websphere-license=True was_url=http://apache-file-server/was-base im_file_name=Install_Mgr_v1.6.2_Lnx_WASv8.5.5.zip sha_im=93e6532d2318d8268313ca62a406a23160306f81 On public cloud deployments (Amazon and Google) the charms are behind security groups or firewalls until the service is exposed. Local deployments have no such restrictions. Please add instructions on how to expose the service to the public at the end of the deployment steps: `juju expose was-base` The readme instructed the user to browse to http://was-base unit's ip-address:9060/ibm/cinsole (should be console) but that is where the guide ends. The console page asks for a user id, and it seems that any user id will work. This was not clear to me when I first did it and should have some explanation in the readme if there is a way to prevent anyone and everyone from logging in. # metadata.yaml This charm only exposes a website relation. The real value of Juju is the ability for charms to relate to each other. The WebSphere products are know for their ability to cluster. I would have expected to see a peers relationship here so if 2 of these servers are deployed they can set up a cluster. Please consider adding a peer relation that sets up WAS clustering, and any other relations that are appropriate. # config-changed and start hook When I increased the instance size I was able to get the charm to install but noticed a bug. WAS deployed without opening any ports on Juju. Since I deployed following the README.md file the was-base charm installed without any values set, so the install, config-changed, and start hook ran (at 15:51:53) but there was no server to start so it exited 0. Then I did the `juju set was-base ...` command (listed above) and the config-changed hook ran successfully (at 16:31:47), but the start hook was not run a second time and never opened the ports. My suggestion is instead of the config-changed hook calling the commands to start WAS call the start hook so the right open-port logic is called. The unit log is attached so you can see the start hook called first and then the install of the software took over 40 minutes. # tests The tests will fail if they don't include constraints for a larger system. Also some comments on what should be in the local.yaml file would be very welcome for someone who didn't write the tests. The configuration for this test needs more than the was_url and the im_file_name, the accept-license ones and the sha_im are needed. Please update the 00-setup to create a valid configuration file so we can test this code. Thanks again for the work here the charm did deploy for me once I increased the size of the amazon instance. I am going to put this back in “in-progress” while these points are addressed. Please put back in Fix Committed when you are ready for another review. As always you can email me with questions or contact us on #juju on chat.freenode.net on IRC.