buttons not working

[Jon Scotto]

I recently upgraded Stash 3.5 to Bitbucket 4.6. I've noticed afterwards that certain buttons (like in creating a pull request do not respond). If i switch to the url that has the 7990 port number, all functions work fine. However the other url we have set as a DNS entry without a port which forwards to the address is missin some functionality.

Another oddity since the upgrade, regardless of what I set as the base url, a message on top in red states that my base url does not match the one I'm connecting to it with. This is incorrect.

[Jon Scotto]

This is our production box. I performed an update last week but haven't been able to find resolution. I'm waiting to hear back from our internal network department to depermine if something needs to be updated with our http url. the http url removes the need to specify the port number. I've advised developers to use the 7990 address until we get this resolved. That seems to work fine so far.

And I wanted to put in a support ticket but the normal link wasn't working at the time.

[Jobin Kuruvilla [Adaptavist]]

Do you have a poxyName set up in your server.xml? If so, check the url. If not, set one with your correct url.

Also checkout https://confluence.atlassian.com/kb/proxying-atlassian-server-applications-with-apache-http-server-806032611.html

[Jon Scotto]

No. I checked the old stash configuration as well. No proxy  is set and both server.xml files appear to match so I'm guessing they are stock configurations.

I did the initial install of Bitbucket on the server, stopped the service, moved the repositories directory over and set the properties files with my Oracle DB settings. This started up without issue. I never found documentation about doing a fresh install using an existing database so this procedure is based on my testings on a non prod instance and non-prod oracle d. 

I just noticed this entry in the log file:

2016-06-28 10:29:41,426 DEBUG [http-nio-7990-exec-5] bitbucketuser @WCH4BTx629x318306x0 svzfqf 10.55.56.95 "POST /admin/server-settings HTTP/1.1" c.a.s.i.e.TransactionAwareEventPublisher Deferring publishing for BaseUrlChangedEvent until AFTER_COMMIT

Could this be a clue? I'm wondering now if the save button on the settings page was not working prior. I tried again to save the setting and it applied this time it seems. The desired url was in the box before and the save appeared to be sucessful. This time however, the red warning bar about the mis-matched baseURL disappeared and I'm getting reports from developers that they are able to perform branching and merging on the .com address (configured at our firewall). the .net address is what i have saved and the system seems to be working fine now.

Thanks for the suggestion anyways

[Josh Wheeler]

Is this on your local machine or on production?