Scaling Activity Failed

Last updated: 2019-11-19 10:30:06

PDF

Problem Description

Scaling activities may fail due to many reasons. If this happens, timely troubleshooting can ensure proper execution of scaling activities.

Problem Analysis

Viewing the Cause

Perform the following steps to view details about the failure:

  1. Click the ID of a failed scaling group in the Scaling group.
  2. Select Scaling Activity to view details, see the figure below:

Troubleshooting According to the Reasons

Troubleshoot the problems based on the cause of failure as instructed below. Common causes of scaling activity failures are as follows:

Causes of Scaling Activity Failures

The causes are classified into the following types:

Troubleshooting

CVM Issues

CvmSoldOut

Cause: The specified resource inventory is insufficient.
Solution: Configure Multi-model in the launch configuration, and select other instance specifications and AZs.

InvalidInstanceType

Cause: The specified instance specification is not available.
Solution: Select an available instance specification in the launch configuration. For more information, see Instance Specifications.

InvalidDiskConfig

Cause: The current system disk does not support the model.
Solution: Check the scaling configuration and modify the system disk type. It is recommended to select premium cloud disks or SSD disks.

QuotiesQuotaNotEnough

Cause: A CVM purchase quota is assigned to each user. For the default quota of pay-as-you-go CVM instances, see Purchase Limits for Pay-as-You-Go CVM Instances.
AS will not be able to provide CVMs if this quota is exceeded.
Solution: Reduce the number of scale-out CVMs or apply to raise the quota by Submitting a Ticket.

InstanceTypeNotFound

Cause: The model defined in the launch configuration is incorrect or not available.
Solution: Change the launch configuration.

Image Issues

ImageNotFound

Cause: The image does not exist or has become invalid.
Solution: Check the corresponding configuration in the launch configuration.

ImageTooLarge

Cause: The system disk capacity is smaller than the image size.
Solution: Check the launch configuration and increase the system disk capacity or use a smaller image.

DpdkVpcOnly

Cause: The selected model type only supports VPC.
Solution: Do not use the model on the basic network. Deselect the model in the launch configuration.

SubnetIPNotEnough

Cause: The number of IPs on the VPC subnet is limited.
Solution: Create a subnet and increase its CIDR.

CBS Issues

DiskSoldOut

Cause: The specified resource inventory is insufficient.
Solution: Enable the default CBS function in the launch configuration.

InvalidDiskConfig

Cause: The current system disk does not support the model instance.
Solution: Check the scaling configuration and modify the system disk type. You are advised to select a high-performance CBS or SSD.

CLB Issues

LbNotFound

Cause: CLB does not exist or has become invalid.
Solution: Check the corresponding configuration in CLB and ensure that the CLB instance associated with the scaling group works properly.

LbListenerNotFound

Cause: The listener in CLB does not exist or has become invalid.
Solution: Check the corresponding configuration in CLB and ensure that the CLB instance associated with the scaling group works properly.

LbLocationNotFound

Cause: The forwarding domain name or URL of the listener in CLB does not exist or has become invalid.
Solution: Check the corresponding configuration in CLB and ensure that the CLB instance associated with the scaling group works properly.

LbNotReady

Cause: A CLB job is ongoing.
Solution: Reduce manual operations for a scaling group during a scaling activity to avoid operation exclusions.

Others

LifecycleHookActionAbandon

Cause: A lifecycle hook for scale-out is configured for the scaling group. It is triggered during scale-out of the scaling group and is finally rejected. As a result, the scaling activity is rolled back, and the scale-out instance is released.
Solution: Check the execution policy of the lifecycle hook.