Firewall rebooting due to tRadTask

Description

This article explains firewall reboot issues due to tRadTask and RadDaemon. Issue occurs when using RADIUS+Local Users on the firewall. This issue has been reported in SonicOS 6.2.5.1-26n. 

Cause

This issue occurs due to two instances of tRadTask running. When one of them exits, it changes the value of a pipe in the code  to -1, which should only be done when the process has completely exited. As another process exists, it crashes SonicOS. SonicOS versions before 6.2.5 had a feature where we would send a halt message to a tRadTask and wait for a minute for it to exit, and if it did not exit, SonicOS would execute a taskDelete to kill it. This feature was removed as sometimes tRadTask would take longer to respond while waiting to Network I/O. 

A sample of the stack trace is shown below: 

04/20/2016 08:52:43.160Reboot due to task suspension04/20/2016 08:52:43.160Task Trace tRadTask: 
tRadTask 8030fbf0 f1f999a0 50 STOP 8030fd3c f1f997d0 ad0003 0 
Stack trace of tRadTask: 
0x812ed3ac -> ($13) 
0x8030fd3c 

Resolution

There is no workaround for this issue except disabling RADIUS on the firewall. However this issue has been fixed in SonicOS 6.2.5.2-27n . Alternatively, a hotfix is available from SonicWall Support.

Related Articles

  • Troubleshooting Network throughput, Latency, and Bandwidth Issues with a SonicWall UTM
    Read More
  • How can I configure a syslog server on a SonicWall firewall?
    Read More
  • Troubleshoot steps if a firewall status is Offline on NSM SaaS
    Read More
not finding your answers?