Linked Confluence page doesn't load in JIRA issue

[Florian Zschocke]

I have linked a JIRA issue in a Confluence page. Usually this would show up in the JIRA issue under "Issue Links". But all I get is:

Page Failed to load

The page is linked and I can click the link and the Confluence page will open. But it doesn't show the page title in the JIRA issue.

This might be related to https://community.atlassian.com/t5/Jira-questions/WstxUnexpectedCharException-when-linking-to-JIRA-issue-in/qaq-p/677239

I have the same setup on a staging system where it works. I have not found any difference so far. Both systems access the same LDAP server fpr user management.

What I have found in the access logs (of the reverse proxy) is the following difference.

Working staging system:

"POST /rpc/xmlrpc?xoauth_requestor_id=u.ser HTTP/1.1" 200 625 "-" "Apache-HttpClient/4.4.1 (Java/1.8.0_151)"

Non-working live system:

"POST /rpc/xmlrpc?xoauth_requestor_id=u.ser HTTP/1.1" 403 928 "-" "Apache-HttpClient/4.4.1 (Java/1.8.0_151)"

[Florian Zschocke · Accepted solution]

While researching another problem and trying out the way of an unproxied connector, I found a message in the application logs about "Confluence remote API disabled." Which was actually a difference between the staging and production system. It was disabled on the production system (to prevent outside access to it).

After turning on the Confluence remote API in Confluence the link is working and the page title shows up as expected in the JIRA issue link.

[Eduard M]

Hello Florian,

The HTTP 403 response means authorisation failed. Can you check if Confluence port is correct in URL, connections are enabled both ways and application link is showing Connected in JIRA?

(http://hostname//plugins/servlet/applinks/listApplicationLinks)

[Florian Zschocke]

Hello Eduard,

the application connection is shown as connected on both sides, JIRA and Confluence. I don't use a port in the URL as both are sitting behind the standard port 443 for HTTPS, using a reverse proxy.

[Florian Zschocke]

Hi Eduard,

I checked the trouble shooting guides and found an article on creating an unproxied application link. I would like to try that as both Confluence and JIRA live on the same server instance.

The article says that I must be able to browse to the unproxied URL. Do you know if this is only true to set up the application link or is that true in general, also for later usage. Of course normal users should not be able to browse the unproxied non-SSL address.

[Eduard M]

Hello Florian,

It's needed while creating the application link because your browser needs to be able to access both unproxied URL's. After that it shouldn't be needed anymore (going by the docs, I haven't tried this myself). You can either do it directly from the server itself or make an allowance to your machine only while setting it up. Afterwards test the linked functionalities to make sure we're not going down the wrong path by using this method. If everything works, then you can block all external access to the unproxied addresses and everything should still work well. 

[Florian Zschocke]

Hello Eduard,

it seems to still be working, the application link, that is, after creating it over an unproxied connector on localhost. The only pain is getting the Application Navigator to behave as its behaviour is rather erratic.

[Eduard M]

Hi Florian.

Right, probably because it uses the URL's from the linked apps. But I think you can add/edit the links in the navigator to use the proxied addresses for your apps?

[Florian Zschocke]

Not for the ones created by the application links. I can only hide them, which I did. But the application navigator in general behaves in an unpredictable way, adding and removing entries as it sees fit. I have it in a state I want it now and will simply keep it that way.

Thank you for your help.