> A practice restore into a spare server shows what a RestoreCord snapshot brings back and how long it takes. A short, repeatable routine for server owners.

Source: https://restorecord.rest/practice-restore

# How to test a Discord backup restore before a crisis

The only honest recovery time is the one you measured. A practice restore into a spare server shows what your snapshot brings back, what it skips and how long it really takes.

[Take a snapshot to test](https://restorecord.com/register)[View pricing](https://restorecord.com/pricing)

Short answer

Create an empty server, invite the same bot and place its role above the roles it must create. Preview the snapshot, restore it into the sandbox, time the run and compare channels, roles, overwrites and messages with your live server.

Owners who rest easy aren't the ones with the most backups. They're the ones who have already restored one. A practice run on a quiet afternoon answers every question you'd otherwise be asking in the middle of an incident.

## Build a sandbox

1. 1Create an empty Discord server that only exists for tests.
2. 2Invite the same bot you use for snapshots.
3. 3Grant it Manage Roles and Manage Channels.
4. 4Drag its role above every role it will need to create.

## Preview first

Open the snapshot in the dashboard previewer. It works like a read-only Discord client. Excluded channels won't appear, and a **padlock** marks channels the bot couldn't read when the snapshot ran. Fix that permission now and the next scheduled run will include the channel.

Snapshot preview — read-onlyAurora Community · March 4, 14:05

AC

Community

announcements

general

support

Voice Lounge

general

Mira14:02

Snapshot captured before the permission redesign.

Kelan14:04

Roles and overwrites look right in the preview.

Aurora Bot14:05

Backup completed — 42 channels, 18 roles, 6,300 messages.

Online — 3

Mira

Kelan

Aurora Bot

## Run the restore

Choose the snapshot, a bot and your sandbox as the destination. Set the messages per channel, up to your plan's limit, and decide whether to clear the destination first. In an empty sandbox that choice barely matters. On a real server, clearing deletes the existing structure, which is exactly why you rehearse here.

Note the start time. Large restores take a while because RestoreCord works within Discord's rate limits. Let the job finish instead of starting a second restore on top of it.

## Check what came back

| Check                        | What good looks like                             |
| ---------------------------- | ------------------------------------------------ |
| Channel order and categories | Matches the live server                          |
| Roles and colours            | All present, in the right order                  |
| Permission overwrites        | Private channels are still private               |
| Messages                     | Recent history present up to the count you chose |
| Images and files             | Present if your attachment mode stored them      |

Restored messages are re-posted, not reinstated. Expect a faithful copy of the conversation rather than the original messages with their authors and timestamps.

## When something is missing

| Symptom                  | Likely cause                                                             |
| ------------------------ | ------------------------------------------------------------------------ |
| A channel is absent      | It was excluded, or the bot couldn't read it                             |
| Roles or channels failed | The bot's role is too low, or Manage Roles or Manage Channels is missing |
| Attachments are gone     | Mode was None, the file was over 8 MB, or the source was already deleted |
| The restore seems stuck  | Discord rate limits are slowing it down, so let it finish                |

## How often to rehearse

A practice restore a few times a year is enough for most servers. Rehearse again after anything that changes the shape of your server: a big role redesign, a new category structure, a switch to a different bot, or a plan change that alters message depth or attachment modes. Keep the sandbox around so the next rehearsal takes minutes to start.

## Write it down

Record how long the restore took and anything that was skipped. That number is your real recovery time, and the list is your fix-it list before the next scheduled run.

## Remember the people

Snapshots rebuild the server, not the member list. See [what a snapshot saves](https://restorecord.rest/what-a-snapshot-captures) for the full recovery order, or read about [member recovery on restorecord.com](https://restorecord.com/member-recovery). Then keep the copies flowing with a [backup schedule](https://restorecord.rest/schedule-discord-backups) and [smart pinning](https://restorecord.rest/snapshot-rotation-and-pinning), or head back to [RestoreCord Snapshots](https://restorecord.rest/).

## A restore rehearsal in four steps

1. 1  
### Prepare a sandbox server  
Create an empty server, invite the same bot and drag its role above the roles it will create.
2. 2  
### Preview the snapshot  
Browse channels and messages in the previewer and note any channel shown with a padlock.
3. 3  
### Run and time the restore  
Restore into the sandbox and write down when it started and when it finished.
4. 4  
### Compare the result  
Check channel order, roles, permission overwrites and a sample of messages against the live server.
[Read the backup getting-started guide](https://restorecord.com/docs/backups/getting-started)

## Restore rehearsal questions

### 

### 

### 

## Know your recovery time before you need it

Take a RestoreCord snapshot today and rehearse the restore while nothing is on fire.

[Get started free](https://restorecord.com/register)[View pricing](https://restorecord.com/pricing)
