Jira server Base URL health check fails

Still need help?

The Atlassian Community is here for you.

Ask the community

Platform notice: Server and Data Center only. This article only applies to Atlassian products on the Server and Data Center platforms.

Support for Server* products ended on February 15th 2024. If you are running a Server product, you can visit the Atlassian Server end of support announcement to review your migration options.

*Except Fisheye and Crucible

Problem

The Base URL Health Check ensures that Jira can access its own Base URL. In Jira 7.1 and above, gadget charts are created using Javascript in the browser instead of server-side with Java. This change resulted in significant performance improvements because the server no longer has to process everything, and the browsers can handle some of it. As a result, Jira now requires that the server can connect to itself, which is different from how it operated previously.

Impact

Gadget titles showing as __MSG_gadget.xxxxxxx, __MSG_gadget.activity., __MSG_gadget.filter, _MSG_gadget.project.title_  and so on, as per the example below:

For example:

Understanding the Results

Icon
Result
What this means

Jira can access itself through the configured Base URL

Your dashboard gadgets should be generated and loaded correctly.

Jira is not able to access itself through the configured Base URL. This is necessary so that dashboard gadgets can be generated successfully. Please verify the current Base URL and, if necessary, review your network configurations to resolve the problem.

Your dashboard gadgets may not be loading correctly. Please follow the diagnosis and resolution steps below to resolve the problem.

Diagnosis

Environment

  • iptables has been configured to forward packets from 80 to 8080 and/or 443 to 8443
  • From the Jira server itself, try curl -v <baseurl> to verify if JIRA is able to communicate with itself.
  • Run SSLPoke from the Jira server itself and see if it returns successfully.
  • Additionally, run the httpclienttest from the Jira server itself to confirm if the SSL configuration is okay.
  • There are application links configured on the instance but they seem to be having issues.
  • As the symptoms for both environments are different, diagnosis and resolution steps are segregated into multiple sections below.
  • Entries have been configured in the hosts file.

When accessing the Dashboard, the gadgets serve a response that is processed by the client (ie the browser) to generate gadgets. While this happens, the server also connects to itself to provide the gadget's translated strings. Jira uses the URL the instance is connected to, so the Jira instance needs to be able to access itself.


Without Usage of Proxy

  1. The user sends a request to Jira on loading the Dashboard
  2. Jira will perform a call towards itself through the gadget URL.
  3. Jira replies with the completed page and gadget's translated strings
With Usage of Proxy


1. The user sends a request to Jira on loading the Dashboard
2. Proxy forwarded this request to Jira
3. Jira will perform a call back to itself through the gadget URL to get the gadget's title. Performing requests to Proxy
4. Proxy forwarded this request back to Jira itself to provide the gadget's translated strings
5. Jira replies gadget's translated strings to the proxy
6. Proxy forwarded this accordingly to Jira
7. Jira replies with the completed page to proxy
8. Proxy forwarded this accordingly to use


If it's not able to do so this error will occur. Some known causes are:

  • Basic or certificate-based auth on a proxy layer that Jira is behind
  • The domain name is not resolvable
  • Jira is served on an IP that's in a separate restricted subnet to the server Jira is running on
  • Jira is served with a self-signed certificate that doesn't exist in its own trust store
  • There are more than 3 redirects between the entry point and the final Jira server (this is tested in the base URL check and also configured in our gadgets).

The recommended best practice is to have Jira served on one domain name only. This can have different IPs (for example internal and external) configured by different DNS, however will only resolve to one domain. For example, https://jira.atlassian.com has an external IP of 131.103.28.11 (mapped to by external DNS) and an internal IP of 192.168.1.2 (mapped with an internal DNS). It has one certificate for that FQDN, and all users access it on that domain only.

 Additionally, there are some environmental-specific causes of this problem. Please see below for further information.


Resolution

Cause 1: 

Jira applications with SSL Configured

The following appears in the atlassian-jira.log file:

2016-02-26 14:22:28,843 http-bio-8443-exec-11 ERROR admin 862x395x1 m32d6k xx.xx.xx.xx /secure/MyJiraHome.jspa [c.a.g.r.internal.http.HttpClientFetcher] Unable to retrieve response
javax.net.ssl.SSLHandshakeException: sun.security.validator.ValidatorException: PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to requested target
	at sun.security.ssl.Alerts.getSSLException(Alerts.java:192)Cause
2022-03-29 23:14:36,116-0700 HealthCheck:thread-1 ERROR ServiceRunner     [c.a.t.j.healthcheck.support.GadgetFeedUrlHealthCheck] An error occurred when performing the Gadget feed URL healthcheck
javax.net.ssl.SSLException: Connection reset
	at sun.security.ssl.Alert.createSSLException(Alert.java:127)
	at sun.security.ssl.TransportContext.fatal(TransportContext.java:324)
	at sun.security.ssl.TransportContext.fatal(TransportContext.java:267)
	at sun.security.ssl.TransportContext.fatal(TransportContext.java:262)
	at sun.security.ssl.SSLTransport.decode(SSLTransport.java:135)

Resolution

  1. Verify that the Root and Intermediate CA certificates were added to the keystore when SSL was configured.
  2. Follow our instructions in our Unable to connect to SSL services due to "PKIX Path Building Failed" error knowledge base article for additional details to resolve the problem.
  3. Restart JIRA (if Data Center, restart all nodes one by one)
  4. Additionally, verify the configuration with httpclienttest as described in the readme of that repository, to gather more detail.


Cause 2: Jira Applications with port forwarding configured

This applies to environments that are using port forwarding, such as iptables:

Cause

If port forwarding has been configured to forward packets from 80 to 8080 and/or 443 to 8443, you may see gadget titles be broken and show as __MSG_gadget.xxxxxxx.  This can happen if the rules are configured only to forward packets from external sources.  

Resolution:

You need to create a rule both for packets coming from external IP sources AND for local packets.

For example, in iptables, it should look like this:

iptables -t nat -I OUTPUT -p tcp -o lo --dport 80 -j REDIRECT --to-ports 8080

For firewalld:

firewall-cmd --permanent --add-masquerade
firewall-cmd --permanent --add-forward-port=port=80:proto=tcp:toport=8080

In Windows, we can use the netsh command, like this:

netsh interface portproxy add v4tov4 listenport=80 connectport=8080


Cause 3: Jira applications with a broken Proxy configuration

This applies to environments configured to use a Web Proxy. This can be determined by checking the "JVM Arguments" in the System Information screen for existing proxy configurations.

Cause

Jira is unable to communicate with itself due to a Web Proxy not being able to correctly resolve the Jira server.

Resolution

Add the Jira server's hostname/ to the -Dhttp.nonProxyHosts and -Dhttps.nonProxyHosts arguments as in Configure an outbound proxy for use in Jira server.

Make sure that the same domain is not being mentioned twice in an argument, e.g. - Dhttps .nonProxyHosts = jira|confluence|jira


Cause 4: Jira configured with Apache in front with BasicAuth password restriction

Cause

Jira is unable to communicate with itself due to a Web Proxy requiring authentication.

Bug ticket raise JRA-61273 - Getting issue details... STATUS

Resolution

Add the server IP itself for access without a password:

Order Allow, Deny
Allow from <hostname>
Satisfy Any

Resolution 2

If using Shibbolethauthentication ,

  1. leave the /rest URLs outside theshibbolethauthentication.
  2. /rest URL must not be shibboleth protected.
  3. Example :
<LocationMatch "^(/rest)/[^/]+">
AuthType None
Require all granted
</LocationMatch>

<LocationMatch "/">
AuthType Shibboleth
ShibRequireSession On
require valid-user
ShibUseEnvironment On
Order allow,deny
Allow from all
</LocationMatch>
DescriptionThe Base URL Health Check verifies if Jira is able to access its own Base URL. Starting in Jira 7.1, gadget charts are generated using Javascript in the browser rather than server-side with Java. This had significant performance improvements as the server was offloading processing to the browsers resulting in less CPU usage. Part of the change modified the way Jira previously operated, and now it is required that the server can connect to itself.
ProductJira
PlatformServer, Data Center


Cause 5: Entries have been configured in the hosts file

Cause

Sometimes after server migrations, some errors are shown when there's no DNS in function and an entry is not properly configured in the hosts file.

HealthCheck:thread-3 ERROR [c.a.t.j.healthcheck.support.GadgetFeedUrlHealthCheck] An error occurred when performing the Gadget feed URL healthcheck
org.apache.http.conn.ConnectTimeoutException: Connect to FQDN:443 [FQDN/OLD_IP] failed: connect timed out

Resolution

Check the hosts file in your server and update the IP:

  • Linux: /etc/hosts
  • Windows: C:\Windows\System32\drivers\etc\hosts

Cause 6: Incorrect Traffic Shaping on Reverse Proxy or LB for REST API calls

Cause

While taking advantage of traffic shaping, customers configure the load balancer to identify all external REST API traffic and send it to a dedicated node. While doing this an incorrect configuration of sending all /rest traffic to a single node will cause background rest calls in Jira pages to be directed to that node. This will cause errors when this dedicated Jira REST API node is not up and running. 

Traffic distribution with Atlassian Data Center

Resolution

While doing traffic shaping only configure M2M REST API Calls to the dedicated node, and ensure end-user API calls continue using the same node that user has logged in initially.


Last modified on Mar 26, 2024

Was this helpful?

Yes
No
Provide feedback about this article
Powered by Confluence and Scroll Viewport.