Overview

While users are reporting constant Internet outages, Kerio Control may face Internet disconnection and network links failures. The dmesg will show the following output:

  e1000e 0000:00:19.0 eth5: Reset adapter unexpectedly

  e1000e 0000:00:19.0 eth5: Detected Hardware Unit Hang

The Kerio Control restart fixes the connection issue for a while, but it cannot be treated as a permanent solution.

This article provides instructions on how to fix such issues from Kerio Control SSH or serial console.

Solution

Such dmesg error indicates compatibilities issues with network adapter (eth5) driver e1000e and Intel hardware. It might be caused by the Active-State Power Management (ASPM) feature, which disables the adapter automatically.

The ifconfig -a command might show the memory interrupts.

Such hardware issues can be fixed by modifying internal settings through the SSH console and BIOS.

Important: Before attempting the fixes, please create a backup of Kerio Control Software, Virtual or Hardware appliance.

1. Access Kerio Control through SSH or serial console (hardware boxes only).
   
2. Disable TSO, GSO and GRO using ethtool for the reported interface (eth5 in this case):
   ethtool -K eth5 gso off gro off tso off
3. Disable TCP checksum offloading:
   ethtool -K eth5 tx off rx off
   
4. Disable Active-State Power Management:
   pcie_aspm=off
   
5. Disable enhanced C1 power-saving state (C1E) in the BIOS under Advanced Power Management configuration.

3rd-party discussions

How to fix “eth0: Detected Hardware Unit Hang” in Debian 9?

e1000e Reset adapter unexpectedly / Detected Hardware Unit Hang

What are CPU "C-states" and how to disable them if needed?