Fixing boot image failures
This article explains how to fix an ECS boot image failure that occurs when you start replication or run a disaster recovery drill using cloud disk asynchronous replication. This guidance also applies if you have manually modified a boot image and cannot access the operating system after a disaster recovery drill or failover.
Symptoms
When you start data replication or preview a disaster recovery drill, the system reports a boot image error:
On the Confirm Start Replication page, the Create Instance row in the Basic Information section displays the warning "Total: 1, Boot image exceptions: 1". Click View Details to see the exception details.
Cause
When replication starts, HDR first creates an ECS instance at the disaster recovery site using a proxy image. It then replicates data from the primary site to the new instance and generates or updates a recovery point at the end of each replication cycle. When you run a disaster recovery drill, HDR uses the proxy image to create a drill ECS instance at the disaster recovery site and replaces its data with the latest recovery point.
HDR does not interact directly with the operating system. If the image for a production ECS instance is upgraded, deactivated, or deleted, HDR may not correctly identify the instance's boot mode and operating system version. This can prevent the target ECS instance from booting correctly during a failover or disaster recovery drill. For best practices on boot modes, see Instance boot mode.
Solution
Automatic repair
During the preview, if the system reports an image issue, click View Details for more information.
In the details dialog box, you can select and configure a suitable, available image for one or more ECS instances. The HDR feature integrates with Cloud Assistant to provide a one-click repair. Click Repair to start the automatic repair. This process runs scripts to detect key information, such as the ECS instance's boot mode, operating system, and version, and then matches the instance with a suitable, available boot image. You can view the details of all operations in the Cloud Assistant console.
In the ECS Boot Image Repair dialog box, if Cloud Assistant is unavailable or detection fails, manually select a boot image and click Confirm to start the repair.
Manual repair
If automatic repair fails to find a suitable image, log on to the ECS instance to confirm its boot mode and operating system version. Then, update the ECS boot image settings accordingly and proceed with replication or the disaster recovery drill.