gitlab-org--gitlab-foss/doc/raketasks/backup_restore.md
Kai Kontio b7316ffdff Update backup_restore.md documentation regarding S3 and IAM profiles.
7f8ef19c41 most likely broke the old configuration format where empty string aws_access_key_id & aws_secret_access_key were still OK. From 9.5 onwards moving backups to S3 using IAM profiles will fail if they are configured.
2017-08-28 11:47:09 +00:00

20 KiB

Backing up and restoring GitLab

backup banner

An application data backup creates an archive file that contains the database, all repositories and all attachments.

You can only restore a backup to exactly the same version and type (CE/EE) of GitLab on which it was created. The best way to migrate your repositories from one server to another is through backup restore.

Backup

GitLab provides a simple command line interface to backup your whole installation, and is flexible enough to fit your needs.

Backup timestamp

Note: In GitLab 9.2 the timestamp format was changed from EPOCH_YYYY_MM_DD to EPOCH_YYYY_MM_DD_GitLab version, for example 1493107454_2017_04_25 would become 1493107454_2017_04_25_9.1.0.

The backup 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, plus the GitLab version. The timestamp is needed if you need to restore GitLab and multiple backups are available.

For example, if the backup name is 1493107454_2017_04_25_9.1.0_gitlab_backup.tar, then the timestamp is 1493107454_2017_04_25_9.1.0.

Creating a backup of the GitLab system

Use this command if you've installed GitLab with the Omnibus package:

sudo gitlab-rake gitlab:backup:create

Use this if you've installed GitLab from source:

sudo -u git -H bundle exec rake gitlab:backup:create RAILS_ENV=production

If you are running GitLab within a Docker container, you can run the backup from the host:

docker exec -t <container name> gitlab-rake gitlab:backup:create

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]

Backup strategy option

Note: Introduced as an option in GitLab 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.

Excluding specific directories from the backup

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

  • db (database)
  • uploads (attachments)
  • repositories (Git repositories data)
  • builds (CI job output logs)
  • artifacts (CI job artifacts)
  • lfs (LFS objects)
  • registry (Container Registry images)
  • 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

Uploading backups to a remote (cloud) storage

Starting with GitLab 7.4 you can let the backup script upload the '.tar' file it creates. It uses the Fog library to perform the upload. In the example below we use Amazon S3 for storage, but Fog also lets you use other storage providers. GitLab imports cloud drivers for AWS, Google, OpenStack Swift, Rackspace and Aliyun as well. A local driver is also available.

For omnibus packages, add the following to /etc/gitlab/gitlab.rb:

gitlab_rails['backup_upload_connection'] = {
  'provider' => 'AWS',
  'region' => 'eu-west-1',
  'aws_access_key_id' => 'AKIAKIAKI',
  'aws_secret_access_key' => 'secret123'
  # If using an IAM Profile, don't configure aws_access_key_id & aws_secret_access_key
  # 'use_iam_profile' => true
}
gitlab_rails['backup_upload_remote_directory'] = 'my.s3.bucket'

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

For installations from source:

  backup:
    # snip
    upload:
      # Fog storage connection settings, see http://fog.io/storage/ .
      connection:
        provider: AWS
        region: eu-west-1
        aws_access_key_id: AKIAKIAKI
        aws_secret_access_key: 'secret123'
        # If using an IAM Profile, leave aws_access_key_id & aws_secret_access_key empty
        # ie. aws_access_key_id: ''
        # use_iam_profile: 'true'
      # The remote 'directory' to store your backups. For S3, this would be the bucket name.
      remote_directory: 'my.s3.bucket'
      # Turns on AWS Server-Side Encryption with Amazon S3-Managed Keys for backups, this is optional
      # encryption: 'AES256'
      # Specifies Amazon S3 storage class to use for backups, this is optional
      # storage_class: 'STANDARD'

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 uploading backups create the following IAM profile, replacing my.s3.bucket with the name of your bucket:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Sid": "Stmt1412062044000",
      "Effect": "Allow",
      "Action": [
        "s3:AbortMultipartUpload",
        "s3:GetBucketAcl",
        "s3:GetBucketLocation",
        "s3:GetObject",
        "s3:GetObjectAcl",
        "s3:ListBucketMultipartUploads",
        "s3:PutObject",
        "s3:PutObjectAcl"
      ],
      "Resource": [
        "arn:aws:s3:::my.s3.bucket/*"
      ]
    },
    {
      "Sid": "Stmt1412062097000",
      "Effect": "Allow",
      "Action": [
        "s3:GetBucketLocation",
        "s3:ListAllMyBuckets"
      ],
      "Resource": [
        "*"
      ]
    },
    {
      "Sid": "Stmt1412062128000",
      "Effect": "Allow",
      "Action": [
        "s3:ListBucket"
      ],
      "Resource": [
        "arn:aws:s3:::my.s3.bucket"
      ]
    }
  ]
}

Uploading to locally mounted shares

You may also send backups to a mounted share (NFS / CIFS / SMB / etc.) by using the Fog Local storage provider. 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:

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:

  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'

Specifying a custom directory for backups

If you want to group your backups you can pass a DIRECTORY environment variable:

sudo gitlab-rake gitlab:backup:create DIRECTORY=daily
sudo gitlab-rake gitlab:backup:create DIRECTORY=weekly

Backup archive permissions

The backup archives created by GitLab (1393513186_2014_02_27_gitlab_backup.tar) will have owner/group git:git and 0600 permissions by default. 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

Storing configuration files

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!

If you use an Omnibus package please see the instructions in the readme to backup your configuration. If you have a cookbook installation there should be a copy of your configuration in Chef. 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.

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.

Configuring cron to make daily backups

Note: The following cron jobs do not backup your GitLab configuration files or SSH host keys.

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

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 as described here.

For installation from source

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:

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

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.

Restore

GitLab provides a simple command line interface to backup your whole installation, and is flexible enough to fit your needs.

The restore prerequisites section includes crucial information. Make sure to read and test the whole restore process at least once before attempting to perform it in a production environment.

You can only restore a backup to exactly the same version and type (CE/EE) of GitLab that you created it on, for example CE 9.1.0.

Restore 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).

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 from source). This file contains the database encryption key, CI secret variables, and 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 authentication enabled and GitLab Runners will lose access to your GitLab server.

Depending on your case, you might want to run the restore command with one or more of the following options:

  • BACKUP=timestamp_of_backup - Required if more than one backup exists. Read what the backup timestamp is about.
  • force=yes - Do not ask if the authorized_keys file should get regenerated.

Restore for installation from source

# Stop processes that are connected to the database
sudo service gitlab stop

bundle exec rake gitlab:backup:restore RAILS_ENV=production

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]

Restore for Omnibus installations

This procedure assumes that:

  • You have installed the exact same version and type (CE/EE) 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.

sudo cp 1493107454_2017_04_25_9.1.0_gitlab_backup.tar /var/opt/gitlab/backups/

Stop the processes that are connected to the database. Leave the rest of GitLab running:

sudo gitlab-ctl stop unicorn
sudo gitlab-ctl stop sidekiq
# Verify
sudo gitlab-ctl status

Next, restore the backup, specifying the timestamp of the backup you wish to restore:

# This command will overwrite the contents of your GitLab database!
sudo gitlab-rake gitlab:backup:restore BACKUP=1493107454_2017_04_25_9.1.0

Restart and check GitLab:

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 version of GitLab, the restore command will abort with an error. Install the correct GitLab version and try again.

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.

Example: Amazon EBS

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.

Example: LVM snapshots + rsync

A GitLab server using omnibus-gitlab, with an LVM logical volume mounted at /var/opt/gitlab. Replicating the /var/opt/gitlab directory using rsync would not be reliable because too many files would change while rsync is running. 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. Now we can have a longer running rsync job which will create a consistent replica on the remote server. 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.

Additional notes

This documentation is for GitLab Community and Enterprise Edition. 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.

Issues are stored in the database. They can't be stored in Git itself.

To migrate your repositories from one server to another with an up-to-date version of GitLab, you can use the import rake task 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.

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
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)

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 trackerhere and here as well as stack overflow.