The application links from Crowd to Confluence, JIRA, Bamboo, or FishEye/Crucible are working, but none from these applications to Crowd.
Recently I tried in Confluence to delete and recreate the invalid application link. Deletion worked, it's recreation failed in some endless loop. After removin the form to input the URL the Web browser showed something like "Consumer cannot be automaticall configured because remote consumer info not found."
Since the servers are in AWS (Amazon Web Services) not all ports can be used for communication. Are there any special requirements?
The versions of the Atlassian´products are up-to-date, except Crowd which is 3.1.2.
The first aspect I would like to clarify is that we don't have application links in Crowd. By application link, I understand you are trying to add an application in Crowd according to the Adding an Application documentation.
As this question was raised in the Fisheye/Crucible product, let's focus the integration of this product with Crowd. Please, if you haven't, follow our Connecting to Crowd documentation in order to have all the detailed instructions to integrate bot applications.
There aren't any special requirements if the hosts reside in AWS. You just need to make sure they are able to communicate with each other.
Some questions to be answered:
* Are you accessing both applications through HTTPS?
* Are the base URLs correctly set in both the instances?
* Are they reachable from the each of their servers via the configured base URL?
* Are you using the same URL while entering the link for the application?
One thing that should be noticed is that crowd URL by default is set to http://localhost:8095/crowd/.
You can also check the Fisheye logs while trying to add Crowd User Management to have a better idea of the error being thrown in the logs. The log files are located in <FISHEYE_INST>/var/log/atlassian-fisheye-yyyy-mm-dd.
Holidays and some employee who could not get updates for our licenses in weeks kept me from working on this problem. Sorry for keeping you waiting!
The licenses for FishEye and Crucible are OK now. I simply wrote them into config.xml since a regular login is not possible (the workaround with the admin hash works and I am logged in as admin).
The documentation at https://confluence.atlassian.com/fisheye/connecting-to-crowd-813695565.html is not correct anymore, the "User Directories" are now found in "User Setting," see attached screenshot.This also shows that some "synchronisation" had happened shortly before, which failed. There is already a hyper-link "Directory Configuration Summary" under "Additional Configuration & Troubleshooting." Choosing that reveals:
Directory ID: 2
Name: Crowd Directory
Created date: 2016-03-24 00:00:00.0
Updated date: 2018-08-23 11:08:54.73
Allowed operations: [UPDATE_USER_ATTRIBUTE, UPDATE_GROUP_ATTRIBUTE]
Implementation class: com.atlassian.crowd.directory.RemoteCrowdDirectory
Encryption type: null
Here the application.name is simply "fisheye" which in Crowd "fisheye/crowd" was configured again. Correcting the issue in Crowd (only "fisheye") fixed the issue for Crowd. I could pass the "Authentication test."
On the side of FishEye/Crucible I get the following as result of a test:That looks already quite fine… Here is the configuration of the application link:Next I'll try to start FishEye/Crusible with regular config.xml file and see whether I can log in.