You need to enable JavaScript to run this app.
Elastic Compute Service

Elastic Compute Service

Copy page
Download PDF
Remote log in issue
What should I do if I cannot remotely connect to the instance?
Copy page
Download PDF
What should I do if I cannot remotely connect to the instance?
If you are unable to log in to or connect to your instance, troubleshoot the issue by checking the items in the following table one by one.​
Troubleshooting approach​
​
Troubleshooting item​
​
Description​
Solution​
Resource status​
Check whether the resource status is normal.​
You can only log in to instances in the Running status.​
Check whether the instance status is Running. For more information, see Viewing instance information.​
If the instance status is Stopped, start the instance and log in to it.​
Login password​
Check login credentials.​
When you log in to an instance, make sure that the login credentials (username, password/key) you enter are the same as those specified during instance creation.​
  • Password: If you log in by using a password, make sure that the password is correct. If you forget your password, you can reset the login password for your instance.​
Login port​
Check whether the security group configuration is correct.​
To log in to a Linux instance, you need to open the default TCP port 22.​
In the ECS console, click the name of the instance you cannot log in to to go to the instance details page, select the "Security Group" tab, and make sure that port 22 is added to the inbound rules of a security group associated with the instance.​
If the port is not added, modify the security group access rules.​
​
Check whether the configuration of the firewall is correct.​
Check whether the internal firewall (iptables, firewalld) of the ECS instance is blocking SSH (port 22).​
If the instance has the firewall enabled and has configured rules to block external access, remote login to the instance may fail.​
Resource overload​
Check whether Internet bandwidth exceeds the upper limit.​
If the Internet bandwidth upper limit is exceeded, you may fail to log in to the instance.​
Check whether Internet bandwidth exceeds the upper limit. For more information, see Checking Internet bandwidth upper limit of instance and Checking Internet bandwidth usage.​
For information about how to increase bandwidth, see Modifying the maximum bandwidth.​
​
Check whether CPU or memory load is too high.​
High CPU or memory usage on an instance may prevent users from logging in to it.​
If you have already created an alert task by using the Cloud Monitor service, the system will automatically send you an alert when CPU or memory usage is too high.​
You can upgrade vCPU and memory by changing the instance type.​
Network configuration​
Check whether an EIP is associated with the instance.​
To use third-party remote login tools such as PuTTY or XShell to log in to an instance, an EIP must be associated with the instance.​
For information about how to view the EIPs associated with an instance, see Viewing instance information.​
If no EIP is associated with the instance, associate one with the instance. For more information, see Associating EIPs.​
​
Check whether the local network works as expected.​
Determine whether the login failure is caused by a local network failure.​
Try logging in to the instance from a local host in another network environment, or switch to a mobile hotspot to test access.​
For local network operator issues, we recommend that you contact the network operator.​
Remote connection tools​
Check whether SSH is enabled on Linux.​
Check whether SSH is disabled inside the Linux instance.​
To connect to Linux via Remote Desktop using the ECS Terminal or third-party tools, enable SSH.​
​
Check whether RDP is enabled in Windows.​
Check whether RDP is disabled inside the Windows instance.​
When you use the ECS Terminal or third-party tools to connect to Windows via Remote Desktop, you need to enable Remote Desktop Protocol (RDP).​
Internal operating system issues​
Check whether the login failure is caused by internal reasons of the instance operating system.​
Determine whether the login failure is caused by file system corruption or similar reasons.​
  • Password injection failed: The most common reason for password retrieval failure is a Cloud-init password injection failure. Submit a ticket to resolve this issue.​
  • After a forced shutdown due to file system corruption: There is a small chance that the file system corruption may persist or reoccur, causing the ECS instance to fail to start again. Try using the fsck command to repair the file system.​
Submit a ticket​
Check whether there are any error messages for remote login.​
To address error messages during remote connection, check the SSH configuration based on the prompt.​
Refer to the troubleshooting documentation to check for relevant solutions. If none of the content matches your error message, go to the console, click "Support > Create ticket" in the top menu bar, and fill in the ticket information to obtain technical support.​
​
If you still cannot log in to the ECS instance after the preceding troubleshooting, submit a ticket to obtain technical support.​
Last updated: 2026.04.13 16:50:03