JIRA datacenter clustering not sucess

[Jennings Liu]

HI ,

I installed jira datacenter following the official guide, after added second node, when I access the jira from URL,  I can see the "system Info" shows that the clustering in off.

but I can see all the shared file path:

Location of Jira Local Home	/var/atlassian/application-data/jira
Location of Jira Shared Home	/data/jira/sharedhome
Location of entityengine.xml	file:/opt/atlassian/jira/atlassian-jira/WEB-INF/classes/entityengine.xml
Location of atlassian-jira.log	/var/atlassian/application-data/jira/log/atlassian-jira.log
Location of indexes	/var/atlassian/application-data/jira/caches/indexes
Location of attachments	/data/jira/sharedhome/data/attachments
Location of backups	/data/jira/sharedhome/export

Add each node have configure cluster.properties under /var/atlassian/application-data/jira/

on node1

jira.node.id = node1
jira.shared.home = /data/jira/sharedhome/

on node2

jira.node.id = node2
jira.shared.home = /data/jira/sharedhome/

and the support page show node is not clustered

Can anyone help with this setting?

[Daniel Wong]

Hi Jennings,

After applying the cluster.properties file on the first node, were you prompted to install a Jira Data Centre license?

[Jennings Liu]

Hi Daniel,

Yes,  you can see the license details.  and I had restarted jira service several times. 

[Daniel Wong]

Is your 2nd node reachable? Have you setup your load balancer to push traffic to your 2nd node? A way to test this is to shut the first node down and bring up the 2nd node to see if you can access it.

[Jennings Liu]

Yes, all these tests are successfull 

[BNOC Engineering]

We are experiencing the same symptoms.

Additionally, we verified that none of the dbo.clusternode database tables are being written to by the nodes, but both are reachable and running.

Each node is otherwise functional but not clustering.

Did you find a resolution?

[BNOC Engineering]

I'm not the OP, but I will share the solution we found for identical symptoms:

Root Cause: A non-breaking whitespace character on the jira.node.id line of the cluster.properties file.

Solution: Ensure the cluster.properties file is UTF-8 encoded with unix line endings and that all whitespaces are breaking spaces (normal spaces).

jira.node.id = dev_node1
jira.shared.home = \\\\org\\share\\vendors\\Jira_Dev\\sharedhome
ehcache.listener.hostName = <ip of current node>
ehcache.listener.port = 40001
ehcache.object.port = 40011

We are not sure where this non-breaking whitespace came from, but we suspect that we copied it from some documentation somewhere when creating this file.

One common tool to inspect this with is NotePad++
Show Spaces and Line Endings: View>Show Symbol>Show All Characters

A breaking whitespace will have an orange dot in its place. A non-breaking whitespace will look like a plain space even in this mode. This makes it a nearly invisible difference.

You need to replace any non-breaking whitespaces until all of your whitespace appears with these orange dots.

Good:

Bad:

Convert to Unix Line Endings: Edit > EOL Conversion > Unix(LF)

Some other solutions that atlassian provided were:

1. Make sure jira.node.id can resolve to an ip address on every node. We did this by adding lines to the host files on our cluster members.
2. Add these lines to the cluster.properties file:

   ehcache.listener.hostName = <ip of current node>
   
   ehcache.listener.port = 40001
   
   ehcache.object.port = 40011

3. Restart the service
4. Check the system info and confirm that it is now clustering:

[Daniel Wong]

That's amazing @BNOC Engineering . Very detailed answer and I'm sure someone will find it useful! Thanks!