Virtual Machines

Products

Virtual Machines

Documentation > Products > Virtual Machines > fatal: mm_request_send: write: Broken pipe Appears during SSH Login

fatal: mm_request_send: write: Broken pipe appears for SSH login

Note: The configuration and instructions of Linux in this article have been tested on the CentOS 6.5 64-bit operating system. For other types and versions of the operating system, configuration may vary. Please refer to the official documentation for the details about relevant operating systems.

Problem Description:

When the client connects to the Linux VMs, the command line or secure log displays an error message as follows:

• fatal: mm_request_send: write: Broken pipe.

• Connection closed by 192.168.0.1.

Problem Causes:

This problem may be caused by the abnormal running of SSH service affected by viruses such as udev-fall.

Solution:

Temporary Recovery:

Take the udev-fall virus as an example, an attempt could be made to temporarily restore the SSH service to normal running as follows:

  1. Enter the system via VNC.

  2. Use the following command to query the virus process:

ps aux | grep udev-fall

root 24486 0.0 0.0 108340 1848 pts/0 S 10:30 0:00 /usr/bin/udev-fall

Description: The number in the second column indicates the process ID (PID).

  1. End the virus process with the following command:

kill -9 *

For example:

kill -9 24486

  1. Use the following command to cancel the automatic running of the virus service:

chkconfig udev-fall off

  1. Use the following command to remove the corresponding commands and self-launch configuration of the virus program:

rm -rf /usr/bin/udev-fall

rm -rf /etc/init.d/udev-fall

  1. Restart the SSH service with the following command and then try to reconnect from the client:

service sshd restart

Reliable Treatment:

It can not determine that whether the system is tampered by the virus or malicious intruder, or has other hidden virus files. Therefore, in order to maintain the server's long-term stable running, it is suggested to restore the server to a normal state through the system disk reset.

If your problem still can not solved, please submit open ticket to us.

Update Time:2019-08-28 21:38:32