Monitoring Hung Threads in WAS

Does anyone currently monitor for hung threads in WebShpere? If so, what do you gain by doing this? What is your best practice?

As you know, WAS can have hung threads that don't always result in an issue. I'm fearful if we setup monitoring, we'd spend a lot of time and resources monitoring a hung thread that doesn't result in an outage.

Just looking for some insight.

Find more posts tagged with

Infor Lawson Technology Group - Discussion

Comments

unknown

We have begun to suspect WAS is hanging and, in some cases, stopping for no reason. All signs are of a typical memory leak, but, we cannot confirm that. As a catch all, we have begun rebooting our server every morning and have setup a monitoring script to alert us of any standard Lawson components which are not running.

lbecker

Haven't experienced an issue like this on AIX. Last had to reboot the server 236 days ago (uptime).

The WebSphere Deployment Manager has been up continuously since the last reboot. We restarted the Node Manager last November. The App Server is restarted nightly at 23:30, which takes 2 to 3 minutes minutes. It has no impact on the Job Scheduler. If there is anybody on at that time, they get used to the routine quickly. It would take about an hour to cycle the server, which isn't acceptable. The Lawson environment process are only cycled as necessary.

As for monitoring, we use a tool called 'Sitescope' which logs in to Lawson every 15 minutes. It has no authority to run any forms, just to login. If it logs in, it's happy. If it gets anything else, it sends an alert. We know WebSphere is down, hung processes get killed if necessary and the App Server is cycled. This seems to happen once every 2 or 3 months.

unknown

We do monitor WAS for hung threads. Our sys admin gets emailed when we see hung threads. We gain information into the number of threads and can monitor them to make sure they clear properly. If we see the number of threads growing it prompts us to proactively review what is happening and if other areas of the system are seeing performance issues based on the hung threads.

This gives us the chance if needed to clear those hung threads. We recycle our Lawson system once a month. We have found that somewhere between 4-6 weeks works as the recycle timeframe to keep everything running smoothly.

VID018_State Navigation Hot Keys.zip

dan-linzy

How does the sys admin get emailed?
Is this a script that you had obtained from Infor/Lawson or was it developed by your resources?

unknown

What version of WAS are you running we had issues until we went to
WebSphere Application Server (WAS)/IBM 7.0.0.23 and have had no issues since

dan-linzy

IBM WebSphere Application Sever V8.5

Copyright © 2025 Infor. All rights reserved.