When running Windows Discovery under heavy load with WinRM enabled, the PowerShell Session Pool may become unresponsive and the Worker Thread is stuck. This may lead to Discovery Schedules timing out, because a probe never ends. Eventually, all Worker Threads on the MID Server become stuck, rendering the MID Server unusable.
Steps to Reproduce
- Set up credentials.
- Set up Discovery with WinRM for Windows. Use IP range that would discover more Windows servers than the number of worker threads.
- Run discovery once.
- Disable all Windows credentials and add a bogus credential.
- Rerun all Windows Classify probes.
The issue is fixed in Kingston Patch 10 and London Patch 2. Please upgrade to one of the fixed releases.
Related Problem: PRB1298206