Phone Lines icon GFI Support Home

Articles in this section

Analyzing Kerio Control crashes

Author: Vladyslav Velychko Updated

Overview

While the Kerio Control is experiencing issues with non-optimized source code, it might crash. The main processes report the major failure and afterward generate the core dump.

If the crash occurred in the Webadmin, the Report Problem window is being displayed with a "Kerio Control has stopped working" message.

CONTROL_-_Report_Problem_Dialog.png

This article provides details on how to gather and analyze such crashing reports.

Prerequisites

Previous Linux experience is highly recommended.

If you cannot follow the instructions below, please open a support ticket.

Information

Kerio Control can generate crash dumps in any of the /var/crash subfolders (depending on the main process responsible for the crash).

var_crash.png

The folder will contain the core.PID file, where PID is the process number (i.e. 1002).

If the crashes occur frequently, it is needed to enable the "Crash-reporting configuration" option in Debug logs and wait for the crash re-occurrence.

crash_reporting_configuration.png

Once the appliance crashes, the logs would show a line similar to:

  • [24/Nov/2022 22:14:13] {assist_config} Process KWF: found new dump /var/crash/winroute/core.1070
    NOTE: the process may be different, depending on the crashing component, and the PID will also be different, the general expected output is core.PID, where PID is a numeric string
  • you will have to navigate into the indicated location and gather the core file for the support team

<supportagent>

NOTE: this entire agent-only section needs reviewing and testing because the core dumps sent via the kassist tool are likely no longer received by Eng, and the GDB utility is not present in the Control boxes by default

Crash reporting mechanism is fully managed by Kerio Assist executable. The process can be initiated manually from the SSH console or serial console (/opt/kerio/winroute folder):

  ./kassist -n "description of the problem" -P KWF -e "your@email.com" -N "location_of_core.PID"

crash_kassist_control.png

As an alternative, you can transfer core.PID file to your local PC using SCP.

Important: the Kerio Control Support is NOT notified for crash reports. It only goes to the Development/QA teams.

The file can be analyzed locally by the GDB debugger tool. The command to be executed from /opt/kerio/winroute folder:

gdb ./winroute <path_to_crash_dump>

Note: /tmp/core.1002 is just an example file here.

gdb1.png

In GDB console, you can execute bt full command to get extended output. Type "quit" to exit the GDB.

gdb2.png

</supportagent>

Provide all this information to the Kerio Control Support team together with Support Information file (in Webadmin Status -> System Health) and dmesg -T output.

Final steps

Once you're sure the file is being transferred to the Support team, you can delete the core file (rm <path_to_core_file>) from your system to reclaim space.

Important: the Box crash might be related to hardware issues. For more information, please refer to Diagnosing Faulty hardware boxes.

Related articles