You can export a new custom image to an OSS bucket and download the image to local computers. This topic describes how to export custom images and what precautions must be taken.
Prerequisites
- An OSS bucket is available within the same region as the custom image to be exported.
If you have not created any OSS buckets, create one. For more information, see Create buckets.
Note When you export a custom image, you are charged for OSS storage and the traffic generated for image download. For more information, see Overview. - The custom image to be exported meets the following requirements:
- It is not created based on an Alibaba Cloud Marketplace image.
- It does not contain a Windows Server operating system.
- It does not contain snapshots of more than four data disks. The size of each data
disk does not exceed 500 GiB.
Note If the capacity of a data disk exceeds 500 GiB, you must distribute the data on the disk to multiple data disks that are smaller than 500 GiB in size and then create and export custom images from these smaller data disks.
Background information
Before you export a custom image, take note of the following points:
- The amount of time it takes to export a custom image depends on the size of the image and the number of concurrent tasks in the queue.
- If an exported custom image contains data disk snapshots, multiple objects appear
in your OSS bucket.
Objects whose names contain system are system disk snapshots. Objects whose names contain data are data disk snapshots. The identifier of a data disk snapshot is the mount point of the source data disk, such as xvdb and xvdc.
- To use the exported image to create identical Linux instances, make sure that the storage location and storage space division of files recorded in /etc/fstab are consistent with the exported data disk snapshot information.
- If the cloud disk does not contain any data when the custom image is created, the decompressed image file also does not contain any data.