Elastic Compute Service:FAQ about O&M and monitoring

• Alibaba Cloud Linux

• CentOS 6, CentOS 7, and CentOS 8 and later

• CoreOS

• Debian 8, Debian 9, and Debian 10 and later

• OpenSUSE

• Red Hat 5, Red Hat 6, and Red Hat 7 and later

  Note

  For Red Hat instances, you must download the Red Hat Package Manager (RPM) package to install Cloud Assistant Agent. For more information, see Install Cloud Assistant Agent.

• SUSE Linux Enterprise Server (SLES) 11, SLES 12, and SLES 15 and later

• Ubuntu 12, Ubuntu 14, Ubuntu 16, and Ubuntu 18 and later

• Windows Server 2012, Windows Server 2016, and Windows Server 2019 and later

The maximum number of Cloud Assistant commands that you can retain in an Alibaba Cloud region varies based on your ECS resource usage. The number can range from 100 to 10,000.

You can modify the name and description of a created Cloud Assistant command, but you cannot modify other information about the command, such as the command content, timeout period, and execution path. If you want to modify the command content or execution path, you can clone the command to create a new command version. For more information, see Clone a command.

Yes, you must install and use Cloud Assistant Agent as an administrator or a root user.

• To run Cloud Assistant commands on ECS instances that run Windows Server operating systems, you must obtain administrator permissions.

• To run Cloud Assistant commands on Linux instances, you must obtain root user permissions.

Yes, you can run a Cloud Assistant command on multiple ECS instances at a time.

• Method 1: In the ECS console, you can select up to 50 instances when you run a command. In an Alibaba Cloud region, you can run Cloud Assistant commands up to 5,000 times per day.

• Method 2: If you want to run a Cloud Assistant command on a large number of ECS instances at a time, we recommend that you use the OOSACS-ECS-BulkyRunCommand template of CloudOps Orchestration Service (OOS). You can use the Tag parameter to specify multiple ECS instances to run the Cloud Assistant command without the need to enter the InstanceId value for each instance. In a single OOS loop task, you can run Cloud Assistant commands up to 1,000 times. You can divide the executions into batches and control the concurrency.

Commands can be run only if all required conditions are met, regardless of whether you run the commands by using Cloud Assistant or after you log on to an ECS instance. To check the execution results, you can use one of the following methods:

• Check the execution results in the ECS console.

• Call the DescribeInvocationResults operation.

• Connect to an ECS instance and view the execution results in log files. For more information, see the How do I view the operation logs of Cloud Assistant? section of this topic.

You failed to install Cloud Assistant Agent because Cloud Assistant Agent is already installed on the Linux instance. To re-install Cloud Assistant Agent, perform the following operations. For more information, see Install Cloud Assistant Agent.

• To re-install Cloud Assistant Agent by using an RPM package, run the following command to forcefully overwrite the existing Cloud Assistant Agent:

  sudo rpm -i --force aliyun_x.x.x.rpm

• To re-install Cloud Assistant Agent by using a Debian package, perform the following steps:

  1. Run the following command to uninstall the existing Cloud Assistant Agent:

     sudo dpkg -r aliyun-assist

  2. Run the following command to re-install Cloud Assistant Agent:

     sudo dpkg -i aliyun_x.x.x.deb

To view the operation logs of Cloud Assistant on an ECS instance, you must connect to the instance. The operation logs of Cloud Assistant are stored in the following files:

• ${install_dir}/${version}/log/aliyun_assist_main.log

• ${install_dir}/${version}/log/aliyun_assist_update.log

By default, Cloud Assistant Agent is installed in the following paths:

• Windows Server: C:\ProgramData\aliyun\assist

• CoreOS: /opt/local/share/aliyun-assist/

• Other Linux operating systems such as Alibaba Cloud Linux, CentOS, Debian, openSUSE, SLES, and Ubuntu: /usr/local/share/aliyun-assist/

Connect to the ECS instance.

• View the status of Cloud Assistant to check whether Cloud Assistant is enabled.

  • Linux:

    Method 1

    sudo systemctl status aliyun

    Method 2

    ps aux | grep aliyun

  • Windows Server:

    Choose Computer Management (Local) > Services and Applications > Services and find Aliyun Service.

• Search for the log file in the directory in which the log file is stored. Example:

  tail -n 100 /usr/local/share/aliyun-assist/$(/usr/sbin/aliyun-service -v)/log/aliyun_assist_main.log

• Search for the heartbeat or update log file in the directory in which the log file is stored. Example:

  tail -n 100 /usr/local/share/aliyun-assist/$(/usr/sbin/aliyun-service -v)/log/aliyun_assist_update.log

You cannot run shell commands in Ubuntu as you run the shell commands in CentOS because the default shell environment of later versions of Ubuntu is dash, not bash. The shell syntax of dash is not fully compatible with bash. To resolve the preceding issue, change the default shell environment of Ubuntu.

1. Run the following command to check whether the shell environment is dash:

   ls /bin/sh

2. If the shell environment is dash, run the following command to switch to the default shell environment:

   dpkg-reconfigure dash

Choose not to specify dash as the default shell environment as prompted.

Yes, ECS instances in the classic network can use Cloud Assistant. We recommend that you install the latest version of Cloud Assistant Agent on the ECS instances. For information about the URL for the latest version of Cloud Assistant Agent, see Install Cloud Assistant Agent. If you installed an earlier version of Cloud Assistant Agent on an ECS instance, you must create a file named region-id in the installation path of Cloud Assistant Agent and enter the region ID of the ECS instance in the file. For information about region IDs, see Regions and zones. For example, perform the following steps for a CentOS instance in the China (Hangzhou) region:

1. Enter the region ID.

   echo 'cn-hangzhou' > /usr/local/share/aliyun-assist/region-id

2. Restart Cloud Assistant.

The message appears because Cloud Assistant does not receive heartbeat information from Cloud Assistant Agent on the ECS instance. To resolve the preceding issue, you can use the following methods:

• If the Cloud Assistant process is disabled, you can restart Cloud Assistant. For information about the solution for Windows Server instances, see the How do I view the status of Cloud Assistant? section of this topic. For Linux instances, perform the following steps:

  1. Run the following command to restart Cloud Assistant:

     sudo systemctl restart aliyun

  2. Run the following command to check whether Cloud Assistant starts on system startup:

     sudo systemctl status aliyun

     If Cloud Assistant does not start on system startup, run the following command to allow Cloud Assistant to start on system startup:

     sudo systemctl enable aliyun

• If the ECS instance resides in the classic network and has an early version of Cloud Assistant Agent installed, you must create a configuration file to specify the region of the instance. For more information, see the Can ECS instances in the classic network use Cloud Assistant? section of this topic.

• Problem description: The operation logs of Cloud Assistant contain the following information:

  2019-11-06 03:10:15,993 INFO  [default] /dev/virtio-ports/org.qemu.guest_agent.0:-1
  2019-11-06 03:10:15,993 ERROR [default] Failed to open gshell: Device or resource busy

• Solutions:

  Method 1

  Run the following command to check whether multiple processes are running on Cloud Assistant Agent. If multiple processes are running, stop all the processes and then restart Cloud Assistant Agent.

  ps aux | grep aliyun

  Method 2

  Run the following command to view which processes are occupying the org.qemu.guest_agent.0 file. All Cloud Assistant processes contain aliyun-** in their names. Stop all non-Cloud Assistant processes.

  lsof /dev/virtio-ports/org.qemu.guest_agent.0

Check whether the environment variables of the Cloud Assistant session contain the commands.

To troubleshoot the issue, check the following items:

• If you use a Windows instance, check whether PowerShell runs as expected on the instance.

• Check whether the task timed out.

• Check whether Cloud Assistant runs as expected on the instance. You can call the DescribeCloudAssistantStatus operation to query the status of Cloud Assistant.

• Check Cloud Assistant logs. By default, Cloud Assistant logs are stored in the following paths:

  • Linux instance: /usr/local/share/aliyun-assist/<Version number of Cloud Assistant Agent>/log/

  • Windows instance: C:\ProgramData\aliyun\assist\<Version number of Cloud Assistant Agent>\log

The Abort state indicates that the task was not sent to the destination instance in 1 minute. We recommend that you resend the task.

If the Abort state persists, check Cloud Assistant logs.

Standard encoding method: By default, commands in Linux instances are UTF-8 encoded and commands in Windows instances are GBK-encoded.

If a Cloud Assistant command contains Chinese characters, make sure that the command is encoded by using the encoding method supported by your ECS instances.

• InstanceNotRunning: The ECS instance is not started. Start the instance.

• ClientNotRunning: Cloud Assistant Agent is not started. Check Cloud Assistant logs to confirm whether Cloud Assistant runs as expected.

• DeliveryTimeout: The Cloud Assistant task failed to be sent. Check Cloud Assistant logs or update Cloud Assistant Agent to the latest version and try again.

The following website O&M suggestions are available:

• Back up disk data on a regular basis. For more information, see Create a snapshot or Enable or disable an automatic snapshot policy.

• Use SSL Certificate Service to authenticate visitors to your website and encrypt data to be transferred over your website. For more information, see What is Certificate Management Service?

• Install a malware scan plug-in and Anti-DDoS service, or enable Security Center. ECS provides free security services. For more information, see Basic security services or Anti-DDoS Basic.

• Monitor the inbound and outbound traffic of the website to identify suspicious traffic. Add Deny security group rules to temporarily control suspicious requests from specific points. For more information, see View the monitoring information of an ECS instance and Add a security group rule.

• Monitor the performance of ECS instances and disks and identify peak traffic periods. Make sure that you are familiar with operations such as configuration upgrade and downgrade, auto scaling, and disk resizing, which you can perform in response to sudden request spikes. For more information, see Overview of instance configuration changes, What is Auto Scaling?, or Overview of disk resizing.

• If you use the root and administrator accounts and the corresponding passwords as credentials to log on to ECS instances, change the passwords on a regular basis. For more information, see Reset the logon password of an instance.

• Update software patches on a regular basis. Alibaba Cloud provides security patches for public images on a regular basis. We recommend that you derive custom images from public images on a regular basis. For more information, see Update a custom image.