PNDA Console is not accessible

Question

After successfully deploying PNDA version 3.4, I cannot find a way to open the PNDA Console. 

I can successfully SSH into the bastion VM, and from there, I can ssh into any VM.

The cdh-edge VM, which is the one supposed to serve the console web application, apparently has no web server running at all. I tried running a "wget localhost" from within cdh-edge and I get connection refused. I tried this in all VMs (and from other VMs, I tried "wget <cdh-edge-ip>" with no success), and I find no web server running.

The were some errors logged when deploying the VMs, I saved a HTML of the stack overview screen, as it is very extensive, you can find it below:

https://drive.google.com/file/d/0B30HPZmaePCHdmctcF9CS2Qtek0/view?usp=sharing

I also included a screenshot of the VMs running, and the CLI output which completed successfully.

Do you have any suggestions on how I can access the PNDA Console?

Thanks in advance

Answer 1

I think there were some issue during deployment, in order to get more details, could you run this command under the pnda-heat-templates folder:

./cli/heat_cli.py -e cation logs

give me the output please so that we will have deeper view on your current issue.

Comments

Thanks for the response Jeremie.

I ran the command successfully, here is the output: http://paste.openstack.org/show/613715/

The errors shown seem to be the same as shown in the dashboard page for this orchestration.

---

ok the error comes from the build environment step, Oracle did update their policy and so the JDK 8u74 is available except if you have an Oracle account. So we recommend to download it manually and add JAVA_HOME in your PATH before building everything.

---

Hi Jeremie, thanks again for your response!

I had already done that in order to build on Release/3.4.

Your comment hinted at the built artifacts, so I decided to check on the package server. It was running, but I ran a tcpdump on it during the deploy and found out no requests were being received due to a network problem. I fixed it and the deployment took a lot longer on the last steps and I finally got to the PNDA Console!

It is unfortunate that there is no indication whatsoever for this kind of problem. After realizing that it was supposed to take a lot longer, I checked the previous log and the deploy_highstate and deploy_orchestrate were indeed too fast, sorta indicating a problem like this.

Thanks for your help in getting PNDA working!

EDIT: I forgot to mention that I also needed to add rules to the edge security group in OpenStack, only port 8888 was allowed, port 80 (which was necessary to access the PNDA Console) was not.