Test results on JIRA Cluster
We have tested JQL Tricks plugin on JIRA Clustered environment using the 6.2 EAP version.
And the result? It works great!
Setup
Two nodes were setup, both on localhost, as explained in the aforementioned document. http://jira.cluster was pointed to the cluster. Aapche load balancing was used and the database was MySQL 5.5.18 Community Server.
Two nodes were setup, both on localhost, as explained in the aforementioned document. http://jira.cluster was pointed to the cluster. Aapche load balancing was used and the database was MySQL 5.5.18 Community Server.
|
Node1
localhost
8080 http://jira1.cluster |
Node 2
localhost
9080 http://jira2.cluster |
Test Results
Searching using the JQLT functions, in general, worked as expected. Following are the key areas which were tested extensively, to verify the behavior in the clustered environment.
|
Scenario
1. Both nodes running 2. Node1 taken down 3. Node1 started back 4. Node2 taken down 5. Node2 started back |
Routing of requests
Requests routed to Node1 Requests routed to Node2 Requests routed to Node2 Requests routed to Node1 Requests routed to Node1 |
Search history
Preserved correctly ✓ Preserved correctly ✓ Preserved correctly ✓ Preserved correctly ✓ Preserved correctly ✓ |
Plugin configurations
Preserved correctly ✓ Preserved correctly ✓ Preserved correctly ✓ Preserved correctly ✓ Preserved correctly ✓ |
JQLT Caching
Worked as expected ✓ Worked as expected ✓ (cache reset and worked as expected for new requests) Worked as expected ✓ Worked as expected ✓ (cache reset and worked as expected for new requests) Worked as expected ✓ |
Conclusion
JQL Tricks plugin works fine in a clustered JIRA environment.
To explain a little bit more on caching, JQLT cache will continue to use EhCache and will be unique per node. When a node goes down, users will be redirected to other nodes in the cluster and the previous cache will be invalidated. It will be treated just like the cache expiry on the same node.
If you have any more questions, please feel free to contact us. If not, enjoy the power of JIRA cluster and advanced JQLT functions!