All Products
Search
Document Center

CloudOps Orchestration Service:ACS-ECS-RescueUnreachableInstance-Linux

Last Updated:Nov 25, 2024

Template name

ACS-ECS-RescueUnreachableInstance-Linux

Execute Now

Template description

When you use an Elastic Compute Service (ECS) instance, it may become unavailable due to several causes. One possible cause is system disk damage, which can occur if the instance is forcibly stopped or restarted, or experiences a sudden breakdown. Other possible causes include failing to update the /etc/fstab file after a data disk is detached from the instance, or loss or corruption of the /etc/fstab or initrd file. Even if the instance becomes unavailable, it may still appear as Running in the ECS console. However, you may experience the following issues: failure to access applications on the instance, disconnection from the network of the instance, and failure to connect to the instance by using Workbench or SSH. If you can establish a Virtual Network Computing (VNC) connection to the instance in the ECS console, you may encounter a message indicating a system startup failure. In this case, you can execute this template to repair the damaged instance. The repair process involves attaching the system disk of the damaged instance to a temporary instance. A repair script is then run on the temporary instance. Finally, the repaired system disk is reattached to the original instance.

Template type

Automated

Owner

Alibaba Cloud

Input parameters

Parameter

Description

Type

Required

Default value

Limit

unreachableInstanceId

The ID of the ECS instance to be repaired.

String

Yes

credentialType

The type of the credential that is used to log on to the ECS instance.

String

Yes

credentialValue

The credential that is used to log on to the ECS instance.

String

Yes

imagePrefix

The name prefix of the image that is used to back up data of the ECS instance.

String

No

OOSRescueBackup-

helperInstanceTypes

The instance types that are available to the temporary ECS instance that is used to repair the damaged ECS instance.

List

No

['ecs.t6-c4m1.large', 'ecs.t5-lc1m1.small', 'ecs.t5-lc1m2.small', 'ecs.s6-c1m1.small', 'ecs.s6-c1m2.small', 'ecs.n1.small', 'ecs.mn4.small', 'ecs.e3.small', 'ecs.e4.small', 'ecs.n2.small', 'ecs.n4.small', 'ecs.t5-lc1m2.large', 'ecs.c6.large', 'ecs.sn2ne.large']

OOSAssumeRole

The Resource Access Management (RAM) role that is assumed by CloudOps Orchestration Service (OOS).

String

No

""

Output parameters

Parameter

Description

Type

diskId

String

imageId

String

rtCommandOutput

String

finalHelperInstanceType

String

Permission policy that is required to execute the template

{
    "Version": "1",
    "Statement": [
        {
            "Action": [
                "ecs:AddTags",
                "ecs:AttachDisk",
                "ecs:CreateImage",
                "ecs:DescribeAvailableResource",
                "ecs:DescribeDisks",
                "ecs:DescribeImages",
                "ecs:DescribeInstances",
                "ecs:DescribeInvocationResults",
                "ecs:DescribeInvocations",
                "ecs:DetachDisk",
                "ecs:RunCommand",
                "ecs:StartInstance",
                "ecs:StopInstance"
            ],
            "Resource": "*",
            "Effect": "Allow"
        },
        {
            "Action": [
                "ros:CreateStack",
                "ros:DeleteStack",
                "ros:GetStack"
            ],
            "Resource": "*",
            "Effect": "Allow"
        }
    ]
}

References

For more information, see ACS-ECS-RescueUnreachableInstance-Linux.yml at GitHub.

Template content

FormatVersion: OOS-2019-06-01
Description:
  en: 'When using ECS instances, some situations can lead to system disk corruption, such as instances being forced to stop or restart, or sudden downtime, failure to update /etc/fstab when the disk is unloaded, or even loss or corruption of /etc/fstab or initrd files. When the instance cannot be accessed, the state of the instance displayed in the ECS instance console may still be Running, but the application in the instance cannot be accessed, the network in the instance cannot be reached, and the connection cannot be established through workbench or SSH. If you can connect to an instance from the console with VNC, you will probably see a page that indicates a system startup failure. At this point, you can consider executing the template to cure the damaged instance. The cure process is that the system disk of the damaged instance will be mounted to the newly created temporary instance, then a cure script will be executed in the temporary instance, and finally the cured system disk will be mounted back to the original instance.'
  zh-cn: the description in Chinese    
  name-en: ACS-ECS-RescueUnreachableInstance-Linux
  name-zh-cn: the description in Chinese
  categories:
    - diagnose
Parameters:
  unreachableInstanceId:
    Label:
      en: UnreachableInstanceId
      zh-cn: the description in Chinese
    Type: String
    AssociationProperty: ALIYUN::ECS::Instance::InstanceId
    AssociationPropertyMetadata:
      RegionId: '{{ ACS::RegionId }}'
  credentialType:
    Description:
      en: 'Credential Type for your unreachable ECS instance after being rescued, either KeyPairName or Password type can be chosen.'
      zh-cn: the description in Chinese
    Label:
      en: CredentialType
      zh-cn: the description in Chinese
    Type: String
    AllowedValues:
      - KeyPairName
      - Password
  credentialValue:
    Description:
      en: 'Credential value for your unreachable ECS instance after being rescued, the value of KeyPairName or Password.'
      zh-cn: the description in Chinese
    Type: String
    Label:
      en: Credential
      zh-cn: the description in Chinese
  imagePrefix:
    Label:
      en: ImagePrefix
      zh-cn: the description in Chinese
    Type: String
    Default: OOSRescueBackup-
  helperInstanceTypes:
    Label:
      en: HelperInstanceTypes
      zh-cn: the description in Chinese
    Type: List
    Default:
      - ecs.t6-c4m1.large
      - ecs.t5-lc1m1.small
      - ecs.t5-lc1m2.small
      - ecs.s6-c1m1.small
      - ecs.s6-c1m2.small
      - ecs.n1.small
      - ecs.mn4.small
      - ecs.e3.small
      - ecs.e4.small
      - ecs.n2.small
      - ecs.n4.small
      - ecs.t5-lc1m2.large
      - ecs.c6.large
      - ecs.sn2ne.large
  OOSAssumeRole:
    Label:
      en: OOSAssumeRole
      zh-cn: the description in Chinese
    Type: String
    Default: ''
RamRole: '{{ OOSAssumeRole }}'
Tasks:
- Name: checkInstanceReady
  Action: 'ACS::CheckFor'
  Description:
    en: Checks ECS instance is linux os
    zh-cn: the description in Chinese
  OnError: ACS::END
  Properties:
    Service: ECS
    API: DescribeInstances
    Parameters:
      InstanceIds:
        - '{{ unreachableInstanceId }}'
    DesiredValues:
      - linux
    PropertySelector: 'Instances.Instance[].OSType'
  Outputs:
    status:
      Type: String
      ValueSelector: 'Instances.Instance[].Status'
    vSwitchId:
      Type: String
      ValueSelector: 'Instances.Instance[].VpcAttributes.VSwitchId'
    zoneId:
      Type: String
      ValueSelector: 'Instances.Instance[].ZoneId'
    oSNameEn:
      Type: String
      ValueSelector: 'Instances.Instance[].OSNameEn'
    oSName:
      Type: String
      ValueSelector: 'Instances.Instance[].OSName'
    imageID:
      Type: String
      ValueSelector: 'Instances.Instance[].ImageId'
- Name: querySystemDisks
  Action: 'ACS::CheckFor'
  Description:
    en: Checks system disk of the ECS instance
    zh-cn: the description in Chinese
  OnError: ACS::END
  Properties:
    Service: ECS
    API: DescribeDisks
    Parameters:
      InstanceId: '{{ unreachableInstanceId }}'
      DiskType: system
    DesiredValues:
      - 'true'
    PropertySelector: '.Disks.Disk[] as $disk|$disk.Category|startswith("cloud") and ($disk.Encrypted|not)|tostring'
  Outputs:
    diskId:
      Type: String
      ValueSelector: 'Disks.Disk[].DiskId'
    category:
      Type: String
      ValueSelector: 'Disks.Disk[].Category'
    encrypted:
      Type: String
      ValueSelector: 'Disks.Disk[].Encrypted'
- Name: whetherStopInstance
  Action: 'ACS::Choice'
  Description:
    en: Choose next task by Instance status
    zh-cn: the description in Chinese
  Properties:
    DefaultTask: stopInstance
    Choices:
      - When:
          'Fn::Equals':
            - Stopped
            - '{{ checkInstanceReady.status }}'
        NextTask: checkAvailableInstanceTypesExist
- Name: stopInstance
  Action: 'ACS::ExecuteAPI'
  Description:
    en: Stops the ECS instances
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: StopInstance
    Parameters:
      InstanceId: '{{ unreachableInstanceId }}'
- Name: Sleep3Minutes
  Description:
    en: Wait instance Stopped
    zh-cn: the description in Chinese
  Action: 'ACS::Sleep'
  Properties:
    Duration: PT3M
- Name: queryUnreachableInstanceStatus
  Action: 'ACS::ExecuteAPI'
  Description:
    en: Query status of unreachable instance
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: DescribeInstances
    Parameters:
      InstanceIds:
        - '{{ unreachableInstanceId }}'
  Outputs:
    status:
      Type: String
      ValueSelector: 'Instances.Instance[].Status'
- Name: whetherForceStopInstance
  Action: 'ACS::Choice'
  Description:
    en: Choose next task by Instance status
    zh-cn: the description in Chinese
  Properties:
    DefaultTask: forceStopInstance
    Choices:
      - When:
          'Fn::Equals':
            - Stopped
            - '{{ queryUnreachableInstanceStatus.status }}'
        NextTask: checkAvailableInstanceTypesExist
- Name: forceStopInstance
  Action: 'ACS::ExecuteAPI'
  Description:
    en: Stops the ECS instances forcibly
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: StopInstance
    Parameters:
      InstanceId: '{{ unreachableInstanceId }}'
      ForceStop: 'true'
- Name: untilStopUnreachableInstanceSuccess
  Action: 'ACS::WaitFor'
  Description:
    en: Waits for the ECS instance to enter stopped status
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: DescribeInstances
    Parameters:
      InstanceIds:
        - '{{ unreachableInstanceId }}'
    DesiredValues:
      - Stopped
    PropertySelector: 'Instances.Instance[].Status'
- Name: checkAvailableInstanceTypesExist
  Action: 'ACS::Template'
  OnError: ACS::END
  Description:
    en: Query current available instance type for creating helper instance in the zone of the unreachable
    zh-cn: the description in Chinese
  Properties:
    TemplateName: 'ACS::ECS::CheckAvailableInstanceTypes'
    Parameters:
      zoneId: '{{ checkInstanceReady.zoneId }}'
      instanceTypes: '{{ helperInstanceTypes }}'
  Outputs:
    availableInstanceType:
      Type: String
      ValueSelector: '.availableInstanceTypes[0]'
- Name: createImage
  Action: 'ACS::ExecuteAPI'
  Description:
    en: Creates a custom image
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: CreateImage
    Parameters:
      ImageName: '{{imagePrefix}}{{ ACS::ExecutionId }}'
      InstanceId: '{{ unreachableInstanceId }}'
      DetectionStrategy: Standard
      Tag:
        - Key: 'instance_to_rescue'
          Value: '{{unreachableInstanceId}}'
        - Key: 'oos_exec'
          Value: '{{ ACS::ExecutionId }}'
  Outputs:
    imageId:
      Type: String
      ValueSelector: ImageId
- Name: createStack
  Action: 'ACS::ExecuteAPI'
  Description:
    en: Create a Ros resource stack
    zh-cn: the description in Chinese
  Properties:
    Service: ROS
    API: CreateStack
    Parameters:
      StackName: 'OOS-{{ACS::ExecutionId}}'
      TimeoutInMinutes: 10
      DisableRollback: false
      Parameters:
        - ParameterKey: helperInstanceType
          ParameterValue: '{{checkAvailableInstanceTypesExist.availableInstanceType}}'
        - ParameterKey: zoneId
          ParameterValue: '{{ checkInstanceReady.zoneId }}'
        - ParameterKey: resourcePrefix
          ParameterValue: 'OOS-{{ACS::ExecutionId}}'
        - ParameterKey: imageId
          ParameterValue: 'centos_8_0_x64_20G_alibase_20191225.vhd'
        - ParameterKey: instanceIdToRescue
          ParameterValue: '{{unreachableInstanceId}}'
        - ParameterKey: executionId
          ParameterValue: '{{ ACS::ExecutionId }}'
      TemplateURL: 'https://oos-debug.oss-cn-hangzhou.aliyuncs.com/ros_template.json'
  Outputs:
    StackId:
      Type: String
      ValueSelector: StackId
- Name: untilImageReady
  Action: ACS::WaitFor
  Description:
    en: Wait for the image to be available
    zh-cn: the description in Chinese
  OnError: deleteStack
  Properties:
    Service: ECS
    API: DescribeImages
    Parameters:
      ImageId: '{{ createImage.imageId }}'
    DesiredValues:
    - Available
    PropertySelector: Images.Image[].Status
  Retries: 50
  Delay: 36
  DelayType: Constant
- Name: untilStackReady
  Action: 'ACS::WaitFor'
  OnError: queryStackStatusReason
  OnSuccess: putRTToHelperInstance
  Description:
    en: Wait for the stack status CREATE_COMPLETE.
    zh-cn: the description in Chinese 
  Properties:
    Service: ROS
    API: GetStack
    Parameters:
      StackId: '{{createStack.StackId}}'
    DesiredValues:
      - CREATE_COMPLETE
    StopRetryValues:
      - CREATE_FAILED
      - CHECK_FAILED
      - ROLLBACK_FAILED
      - ROLLBACK_COMPLETE
      - CREATE_ROLLBACK_COMPLETE
    PropertySelector: Status
  Outputs:
    helperInstanceId:
      Type: String
      ValueSelector: 'Outputs[0].OutputValue'
    statusReason:
      Type: String
      ValueSelector: 'StatusReason'
- Name: queryStackStatusReason
  Action: ACS::ExecuteAPI
  OnError: deleteStack
  OnSuccess: deleteStack
  Description:
    en: Query the reson of failed created stack.
    zh-cn: the description in Chinese 
  Properties:
    Service: ROS
    API: GetStack
    Parameters:
      StackId: '{{createStack.StackId}}'
  Outputs:
    statusReason:
      Type: String
      ValueSelector: 'StatusReason'
- Name: putRTToHelperInstance
  Action: 'ACS::ECS::RunCommand'
  OnError: deleteStack
  Description:
    en: Run cloud assistant command on ECS instance to download rt
    zh-cn: the description in Chinese
  Properties:
    commandContent: 'cd /tmp ; wget https://oos-debug.oss-cn-hangzhou.aliyuncs.com/guestos-scripts-0.0.1.tar.gz; tar -zxvf guestos-scripts-0.0.1.tar.gz'
    commandType: RunShellScript
    instanceId: '{{ untilStackReady.helperInstanceId }}'
- Name: addTags
  Action: ACS::ExecuteAPI
  OnError: deleteStack
  Description:
    en: Add Tags of system disk to instance to rescue
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: AddTags
    Parameters:
      ResourceType: instance
      ResourceId: '{{ unreachableInstanceId }}'
      Tag:
        - Key: 'source_sys_disk'
          Value: '{{ querySystemDisks.diskId }}'
- Name: detachDisk
  Action: 'ACS::ECS::DetachDisk'
  OnError: deleteStack
  Description:
    en: Detaches the system disk from unreachable instance
    zh-cn: the description in Chinese
  Properties:
    instanceId: '{{ unreachableInstanceId }}'
    diskId: '{{ querySystemDisks.diskId }}'
- Name: attachAsDataDisk
  Action: 'ACS::ECS::AttachDisk'
  OnError: deleteStack
  Description:
    en: Attaches the system disk to the helper instance as a data disk
    zh-cn: the description in Chinese
  Properties:
    instanceId: '{{ untilStackReady.helperInstanceId }}'
    diskId: '{{ querySystemDisks.diskId }}'
- Name: runCommand
  Action: 'ACS::ECS::RunCommand'
  OnError: deleteStack
  Description:
    en: Run a cloud assistant command of rescuing disk on ECS instance
    zh-cn: the description in Chinese
  Properties:
    commandContent: cd /tmp/guestos-scripts-0.0.1;./rescue_system_disk.sh
    commandType: RunShellScript
    instanceId: '{{ untilStackReady.helperInstanceId }}'
  Outputs:
    commandOutput:
      Type: String
      ValueSelector: invocationOutput
- Name: forceStopHelperInstance
  Action: 'ACS::ExecuteAPI'
  OnError: deleteStack
  Description:
    en: Stops the helper instance forcibly
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: StopInstance
    Parameters:
      InstanceId: '{{ untilStackReady.helperInstanceId }}'
      ForceStop: 'true'
- Name: untilforceStopHelperInstanceSuccess
  Action: 'ACS::WaitFor'
  OnError: deleteStack
  Description:
    en: Waits for the helper instance to enter stopped status
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: DescribeInstances
    Parameters:
      InstanceIds:
        - '{{ untilStackReady.helperInstanceId }}'
    DesiredValues:
      - Stopped
    PropertySelector: 'Instances.Instance[].Status'
- Name: detachHelperInstanceDataDisk
  Action: 'ACS::ECS::DetachDisk'
  OnError: deleteStack
  Description:
    en: Detaches data disk from the helper instance
    zh-cn: the description in Chinese
  Properties:
    instanceId: '{{ untilStackReady.helperInstanceId }}'
    diskId: '{{ querySystemDisks.diskId }}'
- Name: untilUnreachableInstanceSystemDiskAvailable
  Action: 'ACS::WaitFor'
  OnError: 'ACS::NEXT'
  Description:
    en: Waits for the disk to be detached
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: DescribeDisks
    Parameters:
      DiskIds:
        - '{{ querySystemDisks.diskId }}'
    DesiredValues:
      - Available
    PropertySelector: 'Disks.Disk[].Status'
- Name: deleteStack
  Action: 'ACS::ExecuteApi'
  OnError: 'ACS::NEXT'
  Description:
    en: Delete the ros resource stack
    zh-cn: the description in Chinese
  Properties:
    Service: ROS
    API: DeleteStack
    Parameters:
      StackId: '{{createStack.StackId}}'
- Name: untilStackDeleted
  Action: 'ACS::WaitFor'
  OnError: 'ACS::NEXT'
  Description:
    en: Wait for the ros stack status DELETE_COMPLETE
    zh-cn: the description in Chinese
  Properties:
    Service: ROS
    API: GetStack
    Parameters:
      StackId: '{{createStack.StackId}}'
    DesiredValues:
      - DELETE_COMPLETE
    StopRetryValues:
      - DELETE_FAILED
      - CHECK_FAILED
    PropertySelector: Status
- Name: checkForUnreachableInstanceSystemDiskAvailable
  Action: 'ACS::CheckFor'
  OnError: 'ACS::END'
  Description:
    en: Check for the disk to be detached
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: DescribeDisks
    Parameters:
      DiskIds:
        - '{{ querySystemDisks.diskId }}'
    DesiredValues:
      - Available
    PropertySelector: 'Disks.Disk[].Status'
- Name: whetherCredentialTypeIsKeyPairName
  Action: 'ACS::Choice'
  OnError: 'ACS::NEXT'
  Description:
    en: Choose next task by credential type input
    zh-cn: the description in Chinese
  Properties:
    DefaultTask: attachAsSysDiskWithKeyPairName
    Choices:
      - When:
          'Fn::Equals':
            - Password
            - '{{ credentialType }}'
        NextTask: attachAsSysDisk
- Name: attachAsSysDiskWithKeyPairName
  Action: 'ACS::ExecuteAPI'
  OnSuccess: untilDiskAttached
  OnError: 'ACS::NEXT'
  Description:
    en: Attaches the source system disk to unreachable instance and set PairName credential type for root
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: AttachDisk
    Parameters:
      DiskId: '{{ querySystemDisks.diskId }}'
      InstanceId: '{{ unreachableInstanceId }}'
      Bootable: 'true'
      KeyPairName: '{{credentialValue}}'
- Name: attachAsSysDisk
  Action: 'ACS::ExecuteAPI'
  OnError: 'ACS::NEXT'
  Description:
    en: Attaches the source system disk to unreachable instance and set Password credential type for root
    zh-cn: the description in Chinese
  Properties:
    Service: ECS
    API: AttachDisk
    Parameters:
      DiskId: '{{ querySystemDisks.diskId }}'
      InstanceId: '{{ unreachableInstanceId }}'
      Bootable: 'true'
      Password: '{{credentialValue}}'
- Name: untilDiskAttached
  Action: 'ACS::WaitFor'
  OnError: 'ACS::NEXT'
  Description:
    en: Waits for the system disk to be attached
    zh-cn: the description in Chinese
  Retries: 7
  Properties:
    Service: ECS
    API: DescribeDisks
    Parameters:
      DiskIds:
        - '{{ querySystemDisks.diskId }}'
    DesiredValues:
      - In_use
    PropertySelector: 'Disks.Disk[].Status'
- Name: whetherStartUnreachableInstance
  Action: 'ACS::Choice'
  OnError: 'ACS::NEXT'
  Description:
    en: Choose next task by original instance status
    zh-cn: the description in Chinese
  Properties:
    DefaultTask: ACS::END
    Choices:
      - When:
          'Fn::Equals':
            - Running
            - '{{ checkInstanceReady.status }}'
        NextTask: startUnreachableInstance
- Name: startUnreachableInstance
  Action: 'ACS::ECS::StartInstance'
  Description:
    en: Starts the unreachable instance
    zh-cn: the description in Chinese
  Properties:
    instanceId: '{{ unreachableInstanceId}}'
Outputs:
  diskId:
    Type: String
    Value: '{{ querySystemDisks.diskId }}'
  imageId:
    Type: String
    Value: '{{ createImage.imageId }}'
  rtCommandOutput:
    Type: String
    Value: '{{ runcommand.commandOutput }}'
  finalHelperInstanceType:
    Type: String
    Value: '{{checkAvailableInstanceTypesExist.availableInstanceType}}'
Metadata:
  ALIYUN::OOS::Interface:
    ParameterGroups:
      - Parameters:
          - credentialType
          - credentialValue
          - imagePrefix
          - helperInstanceTypes
        Label:
          default:
            zh-cn: the description in Chinese
            en: Configure Parameters
      - Parameters:
          - unreachableInstanceId
        Label:
          default:
            zh-cn: the description in Chinese
            en: Select ECS Instance
      - Parameters:
          - OOSAssumeRole
        Label:
          default:
            zh-cn: the description in Chinese
            en: Control Options