The following article outlines a short checklist of questions to ask and steps to take before opening a performance ticket. By taking a moment to go through this checklist before opening a request, you're ensuring the fastest possible resolution for the particular issue.
Resolution
For each ticket, please attach your portal-ext.properties file as well as the output of the patching-tool info command specific to the environment.
For performance tickets specifically, please work through the following checklist and include the resulting information when you create a ticket:
Out of Memory
- Are your servers clustered? If so, how many nodes received an OOM?
- If the server goes down:
- Full logs when the production server went down
- Application type/version, Database type/version, OS type/version, JDK type/version, Liferay version, portal.properties
- Cluster environment details and relevant files
- How many days has your server been up before this production down incident?
- What changed in your environment between production up and production down?
- Instructions how to reproduce the production down incident.
- Provide us with Heap Dumps and server logs generated when the OOM exception occurs (See How to Take Heap Dumps)
- JVM settings
CPU Deadlock or Spike
- Work through the same questions as above with one modification.
- Rather than heap dumps, provide a series of thread dumps from your system while the slowness is occurring.
Additional Information
Note: The analysis of the performance of a particular installation depends on many factors, not all of them strictly related to Liferay Portal. However, if you have reasons to believe that there is a performance issue in Liferay Portal, please include the analysis you may have done to point out the issue had its root cause in the product so the Support Service can start working based on your analysis.
Liferay offers other kind of services that better suit these requests. Please contact your account manager for more details.