Linking stash to jira using directory service

Anthony Eghobor

So I create a new directory service using jira crowd and I get this error

"Configure Atlassian Crowd Server 

Bitbucket can use Crowd for user management: users, groups and authentication.

Connection test failed. Response from the server:
org.apache.http.conn.ConnectTimeoutException: Connect to jira.url:443 [ jira.url/ipaddress] failed: connect timed out" . I don't even use apache, I use a load balancer and the SSL terminates at the load balancer before sending the traffic to the jira server on 7999. I use the default tomcat server on the instance. But I believe the connection jira.url:443 is not getting to the load balancer.

And then I read I need to create a connection between both servers first, which I tried to do in jira administration --> Applications --> application links and noticed a weird thing. When I put the stash url to connect from , it tells me to confirm I want to connect from the stash url to the old url of the server I have just migrated from. Basically I did a DB migration, but the connection linking is seeing the url of the old server and not of the new server. 

Is there anyway these two issues are linked or are they independent of each other? Please advise on what to do

 

Andy Heinzer

That connect timed out error is bitbucket telling you that it can't connect to that Jira address listed there.  Why, well I can't say for sure without knowing a lot more about your environment.   Both Jira and Bitbucket are using Tomcat to host their website.  Tomcat is an apache application to host java based websites like Jira, Confluence, Bitbucket, etc. 

There are two different kinds of connections here: User server connection and Application links.  One is not dependent on the other, but if the application server can't reach one, it stands to reason that the other is likely failing for the same reason.

In order to setup the user management, I'd recommend reviewing this guide:  Connecting Bitbucket Server to Jira for user management

As for the application links, I would recommend removing these from both applications first.  Then once this is complete, recreate them.  

 

For both of these kinds of connections though:

You can either use the public facing address for both applications (which probably goes through your load balancer)  OR if both Jira and Bitbucket are behind the same firewall/load balancer, it might not be necessary to run the traffic through that load balancer to make this work.  Instead you could follow How to create an unproxied application link.  This would create a new port for each application to be accessible by, at least within your internal network.  From there it tends to be easier to connect these two applications for situation when you can bypass both SSL and load balancers/proxies.