Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
61 changes: 61 additions & 0 deletions cratedb-ee-docs/features/backups/index.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,61 @@
(backups)=
# Backups

```{include} intro.md
```

## Repositories

Snapshots are stored in a snapshot repository, such as an Amazon S3 bucket, an Azure
Blob Storage container or a local file system. Add a repository in the "Repositories"
tab of the Backups page, with the settings your storage requires. The settings
of each repository type are described in
{ref}`CREATE REPOSITORY <crate-reference:sql-create-repo-types>`. The
credentials are stored by CrateDB only, not by the Console.

Removing a repository from the Console doesn't delete the snapshots stored in
it.

## Backup schedules

To add a schedule, fill in:

- **Name**: a name to recognize the schedule.
- **Repository**: the snapshot repository to store the snapshots in.
- **Schedule**: when to take a snapshot, as a
[cron pattern](https://www.ibm.com/docs/en/db2-as-a-service?topic=task-unix-cron-format)
in UTC. For example, `0 * * * *` takes a snapshot every hour.
- **Scope**: the whole cluster, or a list of tables.
- **Retention**: how many snapshots to keep. See Retention below.

You can pause, edit and delete a schedule at any time. Deleting a schedule
doesn't delete its snapshots.

Snapshots taken by a schedule are named after the schedule's ID and the time
the run started, in UTC.

```{include} snapshots.md
:heading-offset: 1
```

To restore a snapshot, select it in the list. See Restore below.

```{include} retention.md
:heading-offset: 1
```

```{include} restore.md
:heading-offset: 1
```

## Good to know

- Each schedule runs one snapshot at a time. If a snapshot is still running
when the schedule is due again, that run is skipped.
- A snapshot keeps running in CrateDB even if the Console restarts. Its run
is picked up again at the schedule's next run.
- Several schedules can take snapshots at the same time, even in the same
repository.
- Up to 3 snapshots can run at the same time. When 3 are already running, the
next scheduled snapshot waits and starts as soon as one of them finishes.
- Pause your schedules before upgrading CrateDB, and resume them afterwards.
6 changes: 6 additions & 0 deletions cratedb-ee-docs/features/backups/intro.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
Backups are snapshots of your cluster, taken on a schedule and stored in a
snapshot repository. A backup schedule takes a snapshot of your cluster,
at the times you set. It also deletes old snapshots, following its retention settings.

Each time a schedule runs, it starts a snapshot and follows it until it ends. The
run shows the snapshot's result: success, partial or failed.
21 changes: 21 additions & 0 deletions cratedb-ee-docs/features/backups/restore.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
# Restore

You can restore a snapshot that succeeded or partially succeeded. From a partial
snapshot, the incomplete parts are left out.

- **Tables**: restore the tables you select. For a table that already exists,
you can skip it or restore it as a copy, under a new name. Existing tables are
never overwritten.
- **Whole cluster**: restore everything the snapshot contains. Only available
for snapshots of the whole cluster.

:::{caution}
Restoring the whole cluster:
- fails if any table of the snapshot already exists. Drop those tables first,
or restore them one by one as copies;
- replaces users, privileges, views, user-defined functions and cluster
settings with the ones in the snapshot. Users created after the snapshot are
deleted. Make sure the user you log in with exists in the snapshot.
:::

Backup schedules skip their runs while a restore is in progress.
17 changes: 17 additions & 0 deletions cratedb-ee-docs/features/backups/retention.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,17 @@
# Retention

After each successful run, the schedule deletes the snapshots it no longer needs
to keep:

- **Keep last**: the number of newest snapshots that are always kept. Required.
- **Keep all for (days)**: also keep every snapshot taken in the last days.
Optional.
- **Keep one per day for (days)**: also keep the first snapshot of each day.
Optional.

For example, keep last `14`, keep all for `1` day and keep one per day for `14`
days keeps every snapshot of the last day, one snapshot per day for two weeks,
and never fewer than 14 snapshots.

A schedule only deletes its own snapshots. Snapshots taken by hand or by other
schedules are never deleted.
12 changes: 12 additions & 0 deletions cratedb-ee-docs/features/backups/snapshots.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,12 @@
# Snapshots

The Snapshots screen lists your cluster's snapshots, with their state, start time,
duration and failures.

A snapshot can be:

- **In progress**: the snapshot is still running.
- **Success**: every table was saved.
- **Partial**: some parts couldn't be saved. The failures show which ones. The
snapshot can still be restored, without those parts.
- **Failed**: nothing usable was saved. The snapshot can't be restored.
1 change: 1 addition & 0 deletions cratedb-ee-docs/features/index.rst
Original file line number Diff line number Diff line change
Expand Up @@ -15,3 +15,4 @@ In addition to that, the following you can also find the following features:

console
automation
backups/index
Loading