backup_restore.md 19.4 KB
Newer Older
Marin Jankovski's avatar
Marin Jankovski committed
1
2
# Backup restore

3
4
![backup banner](backup_hrz.png)

5
6
7
8
9
10
An application data backup creates an archive file that contains the database,
all repositories and all attachments.
This archive will be saved in `backup_path`, which is specified in the
`config/gitlab.yml` file.
The filename will be `[TIMESTAMP]_gitlab_backup.tar`, where `TIMESTAMP`
identifies the time at which each backup was created.
11

12
13
14
> In GitLab 8.15 we changed the timestamp format from `EPOCH` (`1393513186`)
> to `EPOCH_YYYY_MM_DD` (`1393513186_2014_02_27`)

15
16
You can only restore a backup to exactly the same version of GitLab on which it
was created.  The best way to migrate your repositories from one server to
karen Carias's avatar
karen Carias committed
17
another is through backup restore.
18

19
20
To restore a backup, you will also need to restore `/etc/gitlab/gitlab-secrets.json`
(for omnibus packages) or `/home/git/gitlab/.secret` (for installations
21
22
23
from source). This file contains the database encryption key and CI secret
variables used for two-factor authentication. If you fail to restore this
encryption key file along with the application data backup, users with two-factor
24
authentication enabled will lose access to your GitLab server.
25

26
27
28
## Create a backup of the GitLab system

Use this command if you've installed GitLab with the Omnibus package:
29
```
30
sudo gitlab-rake gitlab:backup:create
31
32
33
```
Use this if you've installed GitLab from source:
```
Ben Bodenmiller's avatar
Ben Bodenmiller committed
34
sudo -u git -H bundle exec rake gitlab:backup:create RAILS_ENV=production
35
```
36
37
If you are running GitLab within a Docker container, you can run the backup from the host:
```
38
docker exec -t <container name> gitlab-rake gitlab:backup:create
39
```
40

41
You can specify that portions of the application data be skipped using the
42
43
44
environment variable `SKIP`. You can skip:

- `db` (database)
45
- `uploads` (attachments)
46
- `repositories` (Git repositories data)
47
48
- `builds` (CI job output logs)
- `artifacts` (CI job artifacts)
49
- `lfs` (LFS objects)
50
- `registry` (Container Registry images)
51

52
Separate multiple data types to skip using a comma. For example:
53
54
55
56
57

```
sudo gitlab-rake gitlab:backup:create SKIP=db,uploads
```

58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
Example output:

```
Dumping database tables:
- Dumping table events... [DONE]
- Dumping table issues... [DONE]
- Dumping table keys... [DONE]
- Dumping table merge_requests... [DONE]
- Dumping table milestones... [DONE]
- Dumping table namespaces... [DONE]
- Dumping table notes... [DONE]
- Dumping table projects... [DONE]
- Dumping table protected_branches... [DONE]
- Dumping table schema_migrations... [DONE]
- Dumping table services... [DONE]
- Dumping table snippets... [DONE]
- Dumping table taggings... [DONE]
- Dumping table tags... [DONE]
- Dumping table users... [DONE]
- Dumping table users_projects... [DONE]
- Dumping table web_hooks... [DONE]
- Dumping table wikis... [DONE]
Dumping repositories:
- Dumping repository abcd... [DONE]
Creating backup archive: $TIMESTAMP_gitlab_backup.tar [DONE]
Deleting tmp directories...[DONE]
Deleting old backups... [SKIPPING]
```

87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
## Backup Strategy Option

> **Note:** Introduced as an option in 8.17

The default backup strategy is to essentially stream data from the respective
data locations to the backup using the Linux command `tar` and `gzip`. This works
fine in most cases, but can cause problems when data is rapidly changing.

When data changes while `tar` is reading it, the error `file changed as we read
it` may occur, and will cause the backup process to fail. To combat this, 8.17
introduces a new backup strategy called `copy`. The strategy copies data files
to a temporary location before calling `tar` and `gzip`, avoiding the error.

A side-effect is that the backup process with take up to an additional 1X disk
space. The process does its best to clean up the temporary files at each stage
so the problem doesn't compound, but it could be a considerable change for large
installations. This is why the `copy` strategy is not the default in 8.17.

To use the `copy` strategy instead of the default streaming strategy, specify
`STRATEGY=copy` in the Rake task command. For example,
`sudo gitlab-rake gitlab:backup:create STRATEGY=copy`.

109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
## Exclude specific directories from the backup

You can choose what should be backed up by adding the environment variable `SKIP`.
The available options are:

* `db`
* `uploads` (attachments)
* `repositories`
* `builds` (CI build output logs)
* `artifacts` (CI build artifacts)
* `lfs` (LFS objects)
* `pages` (pages content)

Use a comma to specify several options at the same time:

```
# use this command if you've installed GitLab with the Omnibus package
sudo gitlab-rake gitlab:backup:create SKIP=db,uploads

# if you've installed GitLab from source
sudo -u git -H bundle exec rake gitlab:backup:create SKIP=db,uploads RAILS_ENV=production
```

132
133
## Upload backups to remote (cloud) storage

134
135
Starting with GitLab 7.4 you can let the backup script upload the '.tar' file it creates.
It uses the [Fog library](http://fog.io/) to perform the upload.
136
137
138
In the example below we use Amazon S3 for storage, but Fog also lets you use
[other storage providers](http://fog.io/storage/). GitLab
[imports cloud drivers](https://gitlab.com/gitlab-org/gitlab-ce/blob/30f5b9a5b711b46f1065baf755e413ceced5646b/Gemfile#L88)
139
for AWS, Google, OpenStack Swift and Rackspace as well. A local driver is
140
[also available](#uploading-to-locally-mounted-shares).
141

142
For omnibus packages, add the following to `/etc/gitlab/gitlab.rb`:
143
144
145
146

```ruby
gitlab_rails['backup_upload_connection'] = {
  'provider' => 'AWS',
147
  'region' => 'eu-west-1',
148
149
  'aws_access_key_id' => 'AKIAKIAKI',
  'aws_secret_access_key' => 'secret123'
150
151
152
  # If using an IAM Profile, leave aws_access_key_id & aws_secret_access_key empty
  # ie. 'aws_access_key_id' => '',
  # 'use_iam_profile' => 'true'
153
154
155
156
}
gitlab_rails['backup_upload_remote_directory'] = 'my.s3.bucket'
```

157
158
Make sure to run `sudo gitlab-ctl reconfigure` after editing `/etc/gitlab/gitlab.rb` to reflect the changes.

159
160
161
162
163
164
165
166
167
For installations from source:

```yaml
  backup:
    # snip
    upload:
      # Fog storage connection settings, see http://fog.io/storage/ .
      connection:
        provider: AWS
168
        region: eu-west-1
169
170
        aws_access_key_id: AKIAKIAKI
        aws_secret_access_key: 'secret123'
171
172
173
        # If using an IAM Profile, leave aws_access_key_id & aws_secret_access_key empty
        # ie. aws_access_key_id: ''
        # use_iam_profile: 'true'
174
175
      # The remote 'directory' to store your backups. For S3, this would be the bucket name.
      remote_directory: 'my.s3.bucket'
176
177
      # Turns on AWS Server-Side Encryption with Amazon S3-Managed Keys for backups, this is optional
      # encryption: 'AES256'
178
179
180
181
```

If you are uploading your backups to S3 you will probably want to create a new
IAM user with restricted access rights. To give the upload user access only for
182
uploading backups create the following IAM profile, replacing `my.s3.bucket`
183
184
185
186
with the name of your bucket:

```json
{
187
  "Version": "2012-10-17",
188
189
  "Statement": [
    {
190
      "Sid": "Stmt1412062044000",
191
192
193
194
195
196
197
      "Effect": "Allow",
      "Action": [
        "s3:AbortMultipartUpload",
        "s3:GetBucketAcl",
        "s3:GetBucketLocation",
        "s3:GetObject",
        "s3:GetObjectAcl",
198
        "s3:ListBucketMultipartUploads",
199
200
201
202
203
204
        "s3:PutObject",
        "s3:PutObjectAcl"
      ],
      "Resource": [
        "arn:aws:s3:::my.s3.bucket/*"
      ]
205
    },
206
    {
207
      "Sid": "Stmt1412062097000",
208
209
      "Effect": "Allow",
      "Action": [
210
211
        "s3:GetBucketLocation",
        "s3:ListAllMyBuckets"
212
213
214
215
      ],
      "Resource": [
        "*"
      ]
216
    },
217
    {
218
      "Sid": "Stmt1412062128000",
219
220
221
222
223
224
225
226
227
228
229
230
      "Effect": "Allow",
      "Action": [
        "s3:ListBucket"
      ],
      "Resource": [
        "arn:aws:s3:::my.s3.bucket"
      ]
    }
  ]
}
```

231
232
233
### Uploading to locally mounted shares

You may also send backups to a mounted share (`NFS` / `CIFS` / `SMB` / etc.) by
234
using the Fog [`Local`](https://github.com/fog/fog-local#usage) storage provider.
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
The directory pointed to by the `local_root` key **must** be owned by the `git`
user **when mounted** (mounting with the `uid=` of the `git` user for `CIFS` and
`SMB`) or the user that you are executing the backup tasks under (for omnibus
packages, this is the `git` user).

The `backup_upload_remote_directory` **must** be set in addition to the
`local_root` key. This is the sub directory inside the mounted directory that
backups will be copied to, and will be created if it does not exist. If the
directory that you want to copy the tarballs to is the root of your mounted
directory, just use `.` instead.

For omnibus packages:

```ruby
gitlab_rails['backup_upload_connection'] = {
  :provider => 'Local',
  :local_root => '/mnt/backups'
}

# The directory inside the mounted folder to copy backups to
# Use '.' to store them in the root directory
gitlab_rails['backup_upload_remote_directory'] = 'gitlab_backups'
```

For installations from source:

```yaml
  backup:
    # snip
    upload:
      # Fog storage connection settings, see http://fog.io/storage/ .
      connection:
        provider: Local
        local_root: '/mnt/backups'
      # The directory inside the mounted folder to copy backups to
      # Use '.' to store them in the root directory
      remote_directory: 'gitlab_backups'
```

274
275
## Backup archive permissions

276
277
The backup archives created by GitLab (`1393513186_2014_02_27_gitlab_backup.tar`)
will have owner/group git:git and 0600 permissions by default.
278
279
280
281
282
283
284
285
286
287
288
289
290
291
This is meant to avoid other system users reading GitLab's data.
If you need the backup archives to have different permissions you can use the 'archive_permissions' setting.

```
# In /etc/gitlab/gitlab.rb, for omnibus packages
gitlab_rails['backup_archive_permissions'] = 0644 # Makes the backup archives world-readable
```

```
# In gitlab.yml, for installations from source:
  backup:
    archive_permissions: 0644 # Makes the backup archives world-readable
```

292
293
## Storing configuration files

294
295
296
297
298
299
Please be informed that a backup does not store your configuration
files.  One reason for this is that your database contains encrypted
information for two-factor authentication.  Storing encrypted
information along with its key in the same place defeats the purpose
of using encryption in the first place!

300
If you use an Omnibus package please see the [instructions in the readme to backup your configuration](https://gitlab.com/gitlab-org/omnibus-gitlab/blob/master/README.md#backup-and-restore-omnibus-gitlab-configuration).
301
If you have a cookbook installation there should be a copy of your configuration in Chef.
302
If you installed from source, please consider backing up your `config/secrets.yml` file, `gitlab.yml` file, any SSL keys and certificates, and your [SSH host keys](https://superuser.com/questions/532040/copy-ssh-keys-from-one-server-to-another-server/532079#532079).
303

304
305
306
307
At the very **minimum** you should backup `/etc/gitlab/gitlab.rb` and
`/etc/gitlab/gitlab-secrets.json` (Omnibus), or
`/home/git/gitlab/config/secrets.yml` (source) to preserve your database
encryption key.
308

309
## Restore a previously created backup
310

311
312
You can only restore a backup to exactly the same version of GitLab that you created it on, for example 7.2.1.

313
314
315
316
317
318
319
320
321
### Prerequisites

You need to have a working GitLab installation before you can perform
a restore. This is mainly because the system user performing the
restore actions ('git') is usually not allowed to create or delete
the SQL database it needs to import data into ('gitlabhq_production').
All existing data will be either erased (SQL) or moved to a separate
directory (repositories, uploads).

322
If some or all of your GitLab users are using two-factor authentication (2FA)
323
324
325
326
then you must also make sure to restore `/etc/gitlab/gitlab.rb` and
`/etc/gitlab/gitlab-secrets.json` (Omnibus), or
`/home/git/gitlab/config/secrets.yml` (installations from source). Note that you
need to run `gitlab-ctl reconfigure` after changing `gitlab-secrets.json`.
327

Valery Sizov's avatar
Valery Sizov committed
328
329
### Installation from source

330
```
331
332
333
# Stop processes that are connected to the database
sudo service gitlab stop

334
bundle exec rake gitlab:backup:restore RAILS_ENV=production
335
336
337
338
339
340
```

Options:

```
BACKUP=timestamp_of_backup (required if more than one backup exists)
341
force=yes (do not ask if the authorized_keys file should get regenerated)
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
```

Example output:

```
Unpacking backup... [DONE]
Restoring database tables:
-- create_table("events", {:force=>true})
   -> 0.2231s
[...]
- Loading fixture events...[DONE]
- Loading fixture issues...[DONE]
- Loading fixture keys...[SKIPPING]
- Loading fixture merge_requests...[DONE]
- Loading fixture milestones...[DONE]
- Loading fixture namespaces...[DONE]
- Loading fixture notes...[DONE]
- Loading fixture projects...[DONE]
- Loading fixture protected_branches...[SKIPPING]
- Loading fixture schema_migrations...[DONE]
- Loading fixture services...[SKIPPING]
- Loading fixture snippets...[SKIPPING]
- Loading fixture taggings...[SKIPPING]
- Loading fixture tags...[SKIPPING]
- Loading fixture users...[DONE]
- Loading fixture users_projects...[DONE]
- Loading fixture web_hooks...[SKIPPING]
- Loading fixture wikis...[SKIPPING]
Restoring repositories:
- Restoring repository abcd... [DONE]
Deleting tmp directories...[DONE]
```
374

Valery Sizov's avatar
Valery Sizov committed
375
376
### Omnibus installations

377
This procedure assumes that:
Valery Sizov's avatar
Valery Sizov committed
378

379
380
381
382
383
384
385
386
- You have installed the exact same version of GitLab Omnibus with which the
  backup was created
- You have run `sudo gitlab-ctl reconfigure` at least once
- GitLab is running.  If not, start it using `sudo gitlab-ctl start`.

First make sure your backup tar file is in the backup directory described in the
`gitlab.rb` configuration `gitlab_rails['backup_path']`. The default is
`/var/opt/gitlab/backups`.
Valery Sizov's avatar
Valery Sizov committed
387
388

```shell
389
sudo cp 1393513186_2014_02_27_gitlab_backup.tar /var/opt/gitlab/backups/
Valery Sizov's avatar
Valery Sizov committed
390
391
```

392
393
Stop the processes that are connected to the database.  Leave the rest of GitLab
running:
Valery Sizov's avatar
Valery Sizov committed
394
395
396
397

```shell
sudo gitlab-ctl stop unicorn
sudo gitlab-ctl stop sidekiq
398
399
400
# Verify
sudo gitlab-ctl status
```
Valery Sizov's avatar
Valery Sizov committed
401

402
403
404
405
Next, restore the backup, specifying the timestamp of the backup you wish to
restore:

```shell
Valery Sizov's avatar
Valery Sizov committed
406
# This command will overwrite the contents of your GitLab database!
407
sudo gitlab-rake gitlab:backup:restore BACKUP=1393513186_2014_02_27
408
```
Valery Sizov's avatar
Valery Sizov committed
409

410
411
412
Restart and check GitLab:

```shell
Valery Sizov's avatar
Valery Sizov committed
413
414
415
416
417
sudo gitlab-ctl start
sudo gitlab-rake gitlab:check SANITIZE=true
```

If there is a GitLab version mismatch between your backup tar file and the installed
418
version of GitLab, the restore command will abort with an error. Install the
419
[correct GitLab version](https://about.gitlab.com/downloads/archives/) and try again.
Valery Sizov's avatar
Valery Sizov committed
420

421
## Configure cron to make daily backups
422

Valery Sizov's avatar
Valery Sizov committed
423
### For installation from source:
424
425
426
427
428
429
430
431
432
```
cd /home/git/gitlab
sudo -u git -H editor config/gitlab.yml # Enable keep_time in the backup section to automatically delete old backups
sudo -u git crontab -e # Edit the crontab for the git user
```

Add the following lines at the bottom:

```
433
# Create a full backup of the GitLab repositories and SQL database every day at 4am
434
0 4 * * * cd /home/git/gitlab && PATH=/usr/local/bin:/usr/bin:/bin bundle exec rake gitlab:backup:create RAILS_ENV=production CRON=1
435
```
436
437
438

The `CRON=1` environment setting tells the backup script to suppress all progress output if there are no errors.
This is recommended to reduce cron spam.
Jacob Vosmaer's avatar
Jacob Vosmaer committed
439

Valery Sizov's avatar
Valery Sizov committed
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
### For omnibus installations

To schedule a cron job that backs up your repositories and GitLab metadata, use the root user:

```
sudo su -
crontab -e
```

There, add the following line to schedule the backup for everyday at 2 AM:

```
0 2 * * * /opt/gitlab/bin/gitlab-rake gitlab:backup:create CRON=1
```

You may also want to set a limited lifetime for backups to prevent regular
backups using all your disk space.  To do this add the following lines to
`/etc/gitlab/gitlab.rb` and reconfigure:

```
# limit backup lifetime to 7 days - 604800 seconds
gitlab_rails['backup_keep_time'] = 604800
```

464
465
466
467
468
469
470
Note that the `backup_keep_time` configuration option only manages local
files. GitLab does not automatically prune old files stored in a third-party
object storage (e.g. AWS S3) because the user may not have permission to list
and delete files. We recommend that you configure the appropriate retention
policy for your object storage. For example, you can configure [the S3 backup
policy here as described here](http://stackoverflow.com/questions/37553070/gitlab-omnibus-delete-backup-from-amazon-s3).

Valery Sizov's avatar
Valery Sizov committed
471
472
NOTE: This cron job does not [backup your omnibus-gitlab configuration](#backup-and-restore-omnibus-gitlab-configuration) or [SSH host keys](https://superuser.com/questions/532040/copy-ssh-keys-from-one-server-to-another-server/532079#532079).

Jacob Vosmaer's avatar
Jacob Vosmaer committed
473
474
475
476
477
## Alternative backup strategies

If your GitLab server contains a lot of Git repository data you may find the GitLab backup script to be too slow.
In this case you can consider using filesystem snapshots as part of your backup strategy.

478
Example: Amazon EBS
Jacob Vosmaer's avatar
Jacob Vosmaer committed
479
480
481
482
483
484

> A GitLab server using omnibus-gitlab hosted on Amazon AWS.
> An EBS drive containing an ext4 filesystem is mounted at `/var/opt/gitlab`.
> In this case you could make an application backup by taking an EBS snapshot.
> The backup includes all repositories, uploads and Postgres data.

485
Example: LVM snapshots + rsync
Jacob Vosmaer's avatar
Jacob Vosmaer committed
486
487

> A GitLab server using omnibus-gitlab, with an LVM logical volume mounted at `/var/opt/gitlab`.
488
> Replicating the `/var/opt/gitlab` directory using rsync would not be reliable because too many files would change while rsync is running.
Jacob Vosmaer's avatar
Jacob Vosmaer committed
489
> Instead of rsync-ing `/var/opt/gitlab`, we create a temporary LVM snapshot, which we mount as a read-only filesystem at `/mnt/gitlab_backup`.
490
> Now we can have a longer running rsync job which will create a consistent replica on the remote server.
Jacob Vosmaer's avatar
Jacob Vosmaer committed
491
492
493
494
> The replica includes all repositories, uploads and Postgres data.

If you are running GitLab on a virtualized server you can possibly also create VM snapshots of the entire GitLab server.
It is not uncommon however for a VM snapshot to require you to power down the server, so this approach is probably of limited practical use.
495

496
497
498
499
500
501
502
## Troubleshooting

### Restoring database backup using omnibus packages outputs warnings
If you are using backup restore procedures you might encounter the following warnings:

```
psql:/var/opt/gitlab/backups/db/database.sql:22: ERROR:  must be owner of extension plpgsql
503
504
psql:/var/opt/gitlab/backups/db/database.sql:2931: WARNING:  no privileges could be revoked for "public" (two occurrences)
psql:/var/opt/gitlab/backups/db/database.sql:2933: WARNING:  no privileges were granted for "public" (two occurrences)
505
506
507
508
509
510
511
512
513
514
515
516
517

```

Be advised that, backup is successfully restored in spite of these warnings.

The rake task runs this as the `gitlab` user which does not have the superuser access to the database. When restore is initiated it will also run as `gitlab` user but it will also try to alter the objects it does not have access to.
Those objects have no influence on the database backup/restore but they give this annoying warning.

For more information see similar questions on postgresql issue tracker[here](http://www.postgresql.org/message-id/201110220712.30886.adrian.klaver@gmail.com) and [here](http://www.postgresql.org/message-id/2039.1177339749@sss.pgh.pa.us) as well as [stack overflow](http://stackoverflow.com/questions/4368789/error-must-be-owner-of-language-plpgsql).

## Note
This documentation is for GitLab CE.
We backup GitLab.com and make sure your data is secure, but you can't use these methods to export / backup your data yourself from GitLab.com.
karen Carias's avatar
karen Carias committed
518
519

Issues are stored in the database. They can't be stored in Git itself.
karen Carias's avatar
karen Carias committed
520

karen Carias's avatar
karen Carias committed
521
To migrate your repositories from one server to another with an up-to-date version of
karen Carias's avatar
karen Carias committed
522
523
524
GitLab, you can use the [import rake task](import.md) to do a mass import of the
repository. Note that if you do an import rake task, rather than a backup restore, you
will have all your repositories, but not any other data.