Help Center/ Elastic Volume Service/ Best Practices/ Setting an Appropriate Automatic Snapshot Policy
Updated on 2026-11-13 GMT+08:00

Setting an Appropriate Automatic Snapshot Policy

Scenarios

An automatic snapshot policy can provide automated data protection for EVS disks through a customized execution cycle and a retention duration. A policy supports day-based or weekly automatic snapshot creation at specified times as well as customized retention durations. In practice, there are often no unified standards for configuring snapshot policies. Different service scenarios have significantly different requirements for data protection. If a policy is configured inappropriately, storage resources may be wasted, or no valid recovery points may be available at critical moments, affecting service continuity. This section describes how to properly configure the execution cycle and retention duration of an automatic snapshot policy, helping you effectively control storage costs while ensuring data recovery capabilities.

For more information about automatic snapshot policies, see Automatic Snapshot Overview.

Configurable Parameters for an Automatic Snapshot Policy

An automatic snapshot policy can be adjusted in two dimensions: execution cycle and retention duration.

The execution cycle determines how often and at what times snapshots are created. Two modes are supported:

  • Day based: automatically creates snapshots at specified hours every N days. Multiple hours can be selected, with a minimum interval of 1 hour. This means you can set multiple execution times within a day, for example, at 02:00 and at 14:00 every day. This mode is suitable when continuous protection is required at a fixed interval.
  • Weekly: automatically creates snapshots at specified hours on specified days of each week. You can select multiple days of the week and multiple hours of the day. For example, you can set the policy to create snapshots at 2:00 every Monday and Thursday. This mode is suitable when protection is required on specific workdays.

The retention duration determines how long snapshots are retained. There are also two modes:

  • By time: retains snapshots for a specified number of days after creation and automatically deletes them upon expiration. This mode is suitable for scenarios with a specific retention period.
  • Continuous retention: retains snapshots indefinitely. When the maximum number of automatic snapshots is reached for the disk, the system automatically deletes the oldest one to make room. This mode is suitable for long-term baseline snapshot retention or compliance retention.
    Figure 1 Configurable parameters for an automatic snapshot policy

Recommended: Snapshot Retention Duration Should Cover Snapshot Creation Cycle

If the retention duration is shorter than the execution cycle, a protection gap will exist between the old and new snapshots. Consider the following example: snapshots are created every two days and retained for one day.

Day 1, 02:00: Snapshot S1 is created.

Day 2, 02:00: S1 expires and is deleted.

Day 3, 02:00: Snapshot S2 is created.

From 02:00 on Day 2 to 02:00 on Day 3, the disk is not protected by any snapshot for 24 hours. If data is corrupted or deleted by mistake during this period, it cannot be restored from snapshots.

Figure 2 Snapshot retention duration shorter than snapshot creation cycle

How to Balance Retention Duration and Snapshot Creation Cycle

The retention duration should match the snapshot creation cycle. The core principle is that the retention duration should be at least equal to the snapshot creation cycle (execution interval in days).

Theoretically, when the retention duration equals the creation cycle, the old and new snapshots can be seamlessly connected. If snapshots are created once a day and retained for one day, it takes approximately 35 minutes to create a snapshot. Snapshot S1 expires 24 hours after creation. The creation of snapshot S2 is triggered at the same time on the next day and takes about 35 minutes to complete, effectively replacing S1.

However, there are risks in practice because this seamless connection requires that the time required for each snapshot creation remains consistent. The time required for creating a snapshot fluctuates depending on disk capacity, I/O load, and underlying storage status. If a snapshot creation takes longer and the previous one has expired before the new one is complete, a protection gap will occur.

Configuration suggestions:

  • The retention duration should be at least one day longer than the execution cycle, leaving a safety margin.
  • For large-capacity disks, it is recommended that the retention duration be at least twice the execution cycle to ensure that at least one snapshot from the previous cycle is available, even if a snapshot creation fails or is delayed.

Figure 3 Snapshot creation fluctuation can cause snapshot gaps
Figure 4 Retention tolerance