---
Title: Manage recovery
URL Source: https://company-skill.com/p/ecs/ecs-manage-recovery
Language: en
Description: You want to protect ECS disk data through backups (snapshots), ensure consistency across multiple disks, restore data when needed, or migrate data between instances or regions. This includes setting…
---

# Manage recovery

Part of **Elastic Compute Service (ECS)**. Route queries via `POST https://company-skill.com/api/route`.

## What You Want to Do

You want to protect ECS disk data through backups (snapshots), ensure consistency across multiple disks, restore data when needed, or migrate data between instances or regions. This includes setting up recurring backups, recovering from accidental deletion, or building disaster recovery workflows.

**Typical User Questions**:
- How to configure automatic snapshots for disks?
- Can I restore a disk from a snapshot?
- How to create snapshot-consistent groups across multiple disks?
- What happens if I delete a snapshot?
- How to migrate data using snapshots?

## Decision Tree

Pick the best path for your situation:

- **If** you want to set up recurring backups with a visual interface and manage policies like **Frequency**, **Retention Period**, and **Creation Time** → Use (go to *ecs/ecs-snapshot*)
- **If** you need programmatic control using APIs like **CreateDiskReplicaPair** or manage **DedicatedBlockStorageCluster** with **Authorization: Bearer** tokens and **RetentionDays** parameters → Use API (go to *ecs/ecs-storage*)
- **If** you’ve already restored a snapshot but need to fix file permissions, or must move small files (<32KB after **Base64 encoding**) between running instances using **Cloud Assistant client**, **Run Command**, or **Send File** with an **InvokeId** and **Execution Plan** → Use (go to *ecs/ecs-instance*)
- **Otherwise (default)** → Start with **** — it’s the safest entry point for most backup needs and supports **Snapshot-Consistent Groups** and **Roll Back** operations without code.

## Path Comparison

| Path | Best For | Complexity | Code Required | Automation | Key Fact | Detail Skill |
|------|----------|------------|---------------|------------|----------|-------------|
| Console / Dashboard | low | No | No | Supports **Snapshot-Consistent Groups** for up to 16 disks (≤32 TiB total) | `ecs/guide/ecs-snapshot` |
| API | medium | Yes | Yes | API calls are synchronous and may time out (504) on long operations | `ecs/api/ecs-storage` |
| Console / Dashboard | medium | No | No | File upload limited to 32KB after **Base64 encoding** | `ecs/troubleshooting/ecs-instance` |

## Path Details

### Path 1: Console / Dashboard
**Brief Description**: The ECS console provides a visual interface to manage **Automatic Snapshot Policies**, allowing you to **Apply or Cancel Policy** on one or more disks. You can configure backup **Frequency**, **Retention Period**, and create **Snapshot-Consistent Groups** to ensure multi-disk consistency. The **Roll Back** feature lets you revert a disk to a prior state.

**Key technical facts**:
- Billing: /0.0001~0.002/

- 1632 TiB

### Path 2: API

**Brief Description**: The ECS Storage API enables programmatic management of enterprise-grade snapshot policies and cross-region disaster recovery via **CreateDiskReplicaPair**. You can control **DedicatedBlockStorageCluster** resources and define retention using **RetentionDays**. Authentication uses **Authorization: Bearer $DASHSCOPE_API_KEY**.

**Key technical facts**:
- Billing: API
- Auth method: Authorization: Bearer $DASHSCOPE_API_KEY

- 12288 GiB

### Path 3: Console / Dashboard
**Brief Description**: When direct disk-level recovery isn’t enough, use the **Cloud Assistant client** to run post-restore fixes inside a running instance. You can **Run Command** or **Send File** (with content in **Base64 encoding**) to transfer small payloads. Each operation returns an **InvokeId** and can be scheduled via an **Execution Plan**.

**Key technical facts**:
- Prerequisites: , Cloud Assistant, 

- Cloud Assistant

- WindowsAliyun Assist Service

## FAQ

Q: Which path should I start with?
A: Start with **** if you’re setting up routine backups — it supports **Snapshot-Consistent Groups**, **Roll Back**, and doesn’t require code.

Q: What if I need to back up 20 disks consistently but used the console path?
A: You’ll hit the **16-disk limit** in **Snapshot-Consistent Groups** — the operation will fail. Use the API path only if you can work around this (e.g., split into groups), but note the 32 TiB total cap still applies.

Q: What if I try to migrate a 100KB config file using Cloud Assistant’s **Send File**?
A: The transfer will fail because **Base64 encoding** of a 100KB file exceeds the 32KB payload limit. You’d need to split the file or use alternative methods (not covered here).

Q: Can I use the API path without storing **Authorization: Bearer** keys securely?
A: No — if your environment can’t protect API credentials (e.g., shared dev machine), you risk unauthorized access. In that case, stick to the console.

Q: Does the console path let me change retention for existing snapshots?
A: No — modifying **Retention Period** in **Automatic Snapshot Policies** only affects *new* snapshots. Existing ones retain their original expiry, per the limitations.

Q: If I need real-time progress on snapshot creation, which path should I avoid?
A: Avoid the **API** path — its synchronous design means no callbacks or webhooks; you must poll manually, and long operations may return 504 errors.

Q: Can I use **Cloud Assistant client** to fix a corrupted filesystem after **Roll Back**?
A: Only if the OS is running and the issue is file-level (e.g., wrong permissions). If the filesystem itself is damaged (disk-level), you must stop the instance and use disk attach/repair — Cloud Assistant can’t help there.

## Related queries

configure automatic snapshots, set up disk backup policy, create snapshot-consistent groups, restore disk from snapshot, migrate data using snapshots, how to enable auto snapshots, can I roll back a disk, what happens if I delete a snapshot, automate snapshot lifecycle, cross-region disk replication

---
Part of [Elastic Compute Service (ECS)](https://company-skill.com/p/ecs.md) · https://company-skill.com/llms.txt