My network drops connection with the message "e1000: eth0: e1000_clean_tx_irq: Detected Tx Unit Hang" with Intel 82545/82546 Gigabit Ethernet Controller

Issue

• The server drops network connections with the following messages in /var/log/messsages

  kernel: e1000: eth0: e1000_clean_tx_irq: Detected Tx Unit Hang 
  kernel: Tx Queue <0> 
  kernel: TDH <7a> 
  kernel: TDT <c5> 
  kernel: next_to_use <c5> 
  kernel: next_to_clean <75> 
  kernel: buffer_info[next_to_clean] 
  kernel: time_stamp <102ab452b> 
  kernel: next_to_watch <7c> 
  kernel: jiffies <102ab52ea> 
  kernel: next_to_watch.status <0> 
  kernel: e1000: eth0: e1000_clean_tx_irq: Detected Tx Unit Hang 
  kernel: Tx Queue <0> 
  kernel: TDH <7a> 
  kernel: TDT <c5> 
  kernel: next_to_use <c5> 
  kernel: next_to_clean <75> 
  kernel: buffer_info[next_to_clean] 
  kernel: time_stamp <102ab452b> 
  kernel: next_to_watch <7c> 
  kernel: jiffies <102ab5aba> 
  kernel: next_to_watch.status <0> 
  kernel: NETDEV WATCHDOG: eth0: transmit timed out 
  kernel: e1000: eth0: e1000_watchdog_task: NIC Link is Up 1000 Mbps Full Duplex, 
  Flow Control: RX