ECS file backup is a refined data protection feature provided by CBR, supporting backup and recovery of files or directories within ECS instances.

Use cases

Backup of key configuration files: regularly back up configuration files of applications such as Nginx, Apache, and Tomcat.

Backup of application data and logs: back up business data directories and logs for archiving or auditing.

Backup of database files: back up physical backup files exported by database dumps.

Cross-instance data distribution: uses backup and restore functions to transfer data between different ECS instances.

Backup methods

ECS file backup offers both automatic and manual backup methods. Detailed explanations and differences between the two methods are as follows:

Item	Automatic backup	Manual backup
Trigger method	Automatically triggered based on the backup rules set in the backup strategy.	Manual trigger.
Backup policy & backup schedule	Required.	Not required.
Recovery point name	The initial name is generated automatically by the system and can be modified later. Name prefix: autobackup_.	Supports customization. Name prefix: manualbackup_.
Use cases	Suitable for scenarios that require regular and frequent data protection, such as daily backup protection of ECS instances, enabling quick data recovery in case of unexpected failures causing data loss.	Applicable for irregular scenarios triggered by specific events, allowing quick recovery of ECS instances or cloud disks to their state before changes after operation failure.

Steps

The steps for manually and automatically backing up ECS local files using the CBR Console are as follows.

Manual backup

Automatic backup

Creating backup source.

Create an ECS local file backup source on the CBR console. The backup source is the backup target, which corresponds to the ECS instance. When creating an ECS instance, you must select install cloud assistant Agent on startup.

Manual backup.

Perform a manual backup on the target backup source. After creating the manual backup task, the system automatically generates the corresponding restore point.

Restore data.

When an ECS instance or cloud disk is abnormal, select the target restore point to recover backup data.

When files in an ECS instance are abnormal, you can restore files based on historical backup points. Supports restoring backed-up ECS files to the source instance or a new ECS instance.

Limits

Limit	Description
Backup source	Directory count limit. Each backup task can back up up to 10 directories. Default excluded directories. Default excludes /dev/, /proc/, /sys/, /boot/, /run/, /var/run/, /lost+found/, and the client installation directory. Only one ongoing task (including backup and restore) is allowed per backup source at the same time. New tasks automatically queue, with a maximum queue length of one. Tasks exceeding this limit will fail to create. When one backup source is linked to multiple backup plans, backup directories specified in each backup policy must not overlap.
Operating system	The current version only supports backup clients for mainstream Linux operating systems (X86 & ARM). Windows operating systems are not supported at this time.
Backup plan	The maximum number of backup plans that a single account can add within a single region is 300. The maximum number of backup sources that a single backup plan can associate with is 300.
Backup strategy	The maximum number of backup policies a single account can add within a single region is 300.
Backup limitations	The current version directly ignores pipe files, socket files, character device files, block device files, symbolic links, and hard links during backup and restore. If file size changes during backup, the size of the successfully uploaded object in the backup repository prevails.
Restore restrictions	Restore failures may occur due to insufficient directory or file permissions. Example: Restore the recovery points that have successfully backed up /src/b/, /src/b/1.txt and /src/b/2.txt to /dst/ according to the retain Latest version policy. If /dst/b/ is an immutable directory and its version is newer than /src/b/ in the recovery point, the restore operation will fail when restoring /src/b/1.txt and /src/b/2.txt because /dst/b/ cannot be written to. Handling name conflicts rules are as follows: Restore directory with existing directory: merge directories. Directory items restore into the target directory with the same name. Files existing in the existing directory but not in the restore directory will not be deleted. Restore directory conflicts with existing files (regular files, hard links, soft links, sockets, block devices, character devices, pipe files, etc.): fail immediately. Restore regular files conflicting with existing directories: fail immediately.
Agent Backup	For optimal backup performance, it is recommended that machines running the CBR Agent are equipped with a dual-core or higher 64-bit CPU and at least 8 GB of available memory.

Billing notes

No fees will be charged if no backup operations are performed after activating the CBR service.

Backing up ECS local files will incur storage costs for backup data. For more information, see Billing Notes.

Cloud Backup and Recovery