How to Solve a Hogger or Stuck Problem in Oracle WebLogic Server?
Identifying the Problem
Application thread which is busy and which is waiting long enough to be detected as “stuck” by WebLogic. If a thread doesn’t return in 600 secs, it gets a flag ‘stuck thread’.
If stuck thread getting increase, there would be some serious problem. WLSDM stuck and hogger thread analyzer is making it easier to you aware before the problem occurs.
Analyzing the Stuck Thread
- Go to “notifications > Metric Notification” in WLSDM menu
2. Click “Stuck Threads” button and analyze complete stuck threads
3. Here is complete stack trace showing how to identify the problem.
Below error massage that we captured from WLSDM stuck thread notification page.
“[ACTIVE] ExecuteThread: ‘11’ for queue: ‘weblogic.kernel.Default (self-tuning)’” Id=171 BLOCKED on weblogic.deployment.jms.PooledConnectionFactory@2a15f1f9 owned by “[ACTIVE] ExecuteThread: ‘39’ for queue: ‘weblogic.kernel.Default (self-tuning)’” Id=228 at weblogic.deployment.jms.PooledConnectionFactory.createConnectionInternal(PooledConnectionFactory.java:366) — blocked on weblogic.deployment.jms.PooledConnectionFactory@2a15f1f9 at weblogic.deployment.jms.PooledConnectionFactory.createConnection(PooledConnectionFactory.java:260) at com.bea.wli.sb.transports.jms.JmsOutboundMessageContext.getSession(JmsOutboundMessageContext.java:438) at com.bea.wli.sb.transports.jms.JmsOutboundMessageContext.newMessage(JmsOutboundMessageContext.java:660) at com.bea.wli.sb.transports.jms.JmsOutboundMessageContext.setRequestMetaData(JmsOutboundMessageContext.java:252) wli.sb.transports.jms.JmsOutboundMessageContext.send(JmsOutboundMessageContext.java:587) at com.bea.wli.sb.transports.jms.JmsTransportProvider.sendMessageAsync(JmsTransportProvider.java:771)
4. How to Solve Stuck Thread
Set alarm to “Stuck Thread” metric in WLSDM. When stuck thread increase in WebLogic Server alarm will be triggered. If you have alarm from WLSDM stuck thread metric, follow the below instructions:
· Quick health check of application server.
· Database and other integrated technologies wherever your application reaching like LDAP server for authentication.
· In parallel review thread dumps for stuck threads.
Go to “Smart Dashboards > Servers” then click “Metric & Chart Options” button for define alarm.
Other WLSDM Stuck and Hogger Thread Articles:
- https://wlsdm.com/tutorials/how-to-retrieve-the-current-stack-information-from-a-weblogic-server-continuously
- https://medium.com/wlsdmforweblogic/weblogic-management-weblogic-performance-issues-b9b719beffa5
- https://medium.com/wlsdmforweblogic/analyze-stuck-and-hogging-threads-62ed839be6ce
Installation is really easy and you can setup a complete monitoring infrastructure in less than 5 minutes. If you want to try then go to download page below URL:
