datastore-backup

[Exporting Data from Google Cloud Datastore](https://cloud.google.com/datastore/docs/export-import-entities#rest_ seems easy:

From the Commandline:

sh``` gcloud datastore export gs://exampleBucket/backupdir

Or in node.js:

js```
    const datastore = new Datastore()
    datastore.export({outputUrlPrefix: 'gs://exampleBucket/backupdir'});

So why do you need a library for that? Because it is in the details!

For one thing the code above contains an export containing all kinds can not be loaded into BigQuery: Data exported without specifying an entity filter cannot be loaded into BigQuery."

So you need a list of all Kinds in the Datastore. But there are Limits on entity filters: "Each request is limited to 100 entity filter combinations, where each combination of filtered kind and namespace counts as a separate filter towards this limit." (It was 50 e few years before.)

And there is an other limit: "The managed export and import service limits the number of concurrent exports and imports to 50 and allows a maximum of 20 export and import requests per minute for a project."

So if you have a lot of entities you have to do some slicing and dicing.

CLI Usage

Command line usage is simple. You have to provide the projectId of the datastore and a bucket name. Both have to be in the same region.

% yarn ts-node src/bin/backup.ts --help
usage: backup.ts [-h] [-v] [-d BACKUPDIR] [-n BACKUPNAME] [-s NAMESPACE] projectId bucket

Backup Datastore.

positional arguments:
  projectId             Datastore project ID
  bucket                GCS bucket to store backup

optional arguments:
  -h, --help            show this help message and exit
  -d BACKUPDIR, --backupDir BACKUPDIR
                        prefix/dir within bucket
  -n BACKUPNAME, --backupName BACKUPNAME
                        name of backup (default: autogenerated)
  -s NAMESPACE, --namespace NAMESPACE
                        datastore namespace

Please provide `GOOGLE_APPLICATION_CREDENTIALS` via the Environment!

% export GOOGLE_APPLICATION_CREDENTIALS=~/sampleproj-b0a74af0545e.json
% yarn ts-node src/bin/backup.ts sampleproj samplebucket-tmp
ℹ backup to gs://hudora-tmp/bak/20211221T212651-sampleproj
ℹ Dumping datastore sampleproj
✔ 2 Kinds
2 kinds  ℹ Dumping NumberingAncestor, NumberingItem
2 kinds  ℹ Dumping to gs://samplebucket-tmp/bak/20211221T212651-sampleproj-0
2 kinds  ✔ Dumping finished 27150 records (4.04 MB) in 51s
2 kinds  ✔ written gs://samplebucket-tmp/bak/20211221T212651-sampleproj-0/20211221T212651-sampleproj-0.overall_export_metadata
✨  Done in 54.16s.

This generates a single folder structure under gs://samplebucket-tmp/bak/20211221T212651-sampleproj-0. For projects with more than 100 entities more folders will be generated.

There is one subfolder per kind containing the Entities encoded in LevelDB Format.

Programmatic Usage

import { Datastore } from '@google-cloud/datastore';
import { dumpAllKinds } from '../lib/datastore-backup';

await dumpAllKinds(new Datastore({ projectId: 'sampleproj' }), 'samplebucket-tmp')

See also

• Google DatastoreImport/Export Documentation
• node.js low-level Datastore export API
• firestore-to-bigquery-export npm module for Cloud Functions
• firestore-to-bigquery-export npm module an other one wth some documentation

Changelog

1.0.8 (2021-12-21)

Changelog

All notable changes to this project will be documented in this file. See standard-version for commit guidelines.