What should I do if the error "fatal: mm_request_send: write: Broken pipe" is displayed?

Problem Phenomenon

When executing the SSH command in the ECS instance of the Linux system, the error "fatal: mm_request_send: write: Broken pipe" is displayed.

Problem Causes

This problem may be caused by viruses such as udev-fall that cause the SSH service to run abnormally.

Solution

You can solve this problem through a temporary solution (terminating the virus process) or a complete solution (using a snapshot to roll back the cloud disk).

Method 1: Temporary processing method

This article takes the udev-fall virus as an example to temporarily restore the normal operation of the SSH service.

1. Connect to the Linux instance through VNC.

2. Execute the following command to check the udev-fall virus process information and record the process ID.

ps aux | grep udev-fall

3. Execute the following command to end the udev-fall virus process.

kill -9 [$PID] # [$PID] indicates the udev-fall virus process ID obtained in the previous step.

4. Execute the following command to cancel the automatic running setting of the udev-fall virus program.

chkconfig udev-fall off

5. Execute the following command to delete all udev-fall virus program-related instructions and startup configurations.

for i in `find / -name "udev-fall"`;
do echo '' > $i && rm -rf $i;
don't

6. Run the following command to restart the SSH service.

systemctl restart sshd.service

Method 2: Complete solution

Since it is impossible to know whether viruses or malicious intruders have tampered with the system or hidden other virus files, for the long-term stable operation of the server, it is recommended that you restore the server to a normal state by rolling back the cloud disk snapshot.