debian-mirror-gitlab/doc/development/migration_style_guide.md

Migration Style Guide

When writing migrations for GitLab, you have to take into account that these will be run by hundreds of thousands of organizations of all sizes, some with many years of data in their database.

In addition, having to take a server offline for an upgrade small or big is a big burden for most organizations. For this reason, it is important that your migrations are written carefully, can be applied online, and adhere to the style guide below.

Migrations are not allowed to require GitLab installations to be taken offline unless absolutely necessary.

When downtime is necessary the migration has to be approved by:

1. The VP of Engineering
2. A Backend Maintainer
3. A Database Maintainer

An up-to-date list of people holding these titles can be found at https://about.gitlab.com/company/team/.

When writing your migrations, also consider that databases might have stale data or inconsistencies and guard for that. Try to make as few assumptions as possible about the state of the database.

Please don't depend on GitLab-specific code since it can change in future versions. If needed copy-paste GitLab code into the migration to make it forward compatible.

For GitLab.com, please take into consideration that regular migrations (under db/migrate) are run before Canary is deployed, and post-deployment migrations (db/post_migrate) are run after the deployment to production has finished.

Schema Changes

Changes to the schema should be committed to db/structure.sql. This file is automatically generated by Rails, so you normally should not edit this file by hand. If your migration is adding a column to a table, that column will be added at the bottom. Please do not reorder columns manually for existing tables as this will cause confusing to other people using db/structure.sql generated by Rails.

When your local database in your GDK is diverging from the schema from master it might be hard to cleanly commit the schema changes to Git. In that case you can use the scripts/regenerate-schema script to regenerate a clean db/structure.sql for the migrations you're adding. This script will apply all migrations found in db/migrate or db/post_migrate, so if there are any migrations you don't want to commit to the schema, rename or remove them. If your branch is not targeting master you can set the TARGET environment variable.

# Regenerate schema against `master`
scripts/regenerate-schema

# Regenerate schema against `12-9-stable-ee`
TARGET=12-9-stable-ee scripts/regenerate-schema

What Requires Downtime?

The document "What Requires Downtime?" specifies various database operations, such as

• dropping and renaming columns
• changing column constraints and types
• adding and dropping indexes, tables, and foreign keys

and whether they require downtime and how to work around that whenever possible.

Downtime Tagging

Every migration must specify if it requires downtime or not, and if it should require downtime it must also specify a reason for this. This is required even if 99% of the migrations won't require downtime as this makes it easier to find the migrations that do require downtime.

To tag a migration, add the following two constants to the migration class' body:

• DOWNTIME: a boolean that when set to true indicates the migration requires downtime.
• DOWNTIME_REASON: a String containing the reason for the migration requiring downtime. This constant must be set when DOWNTIME is set to true.

For example:

class MyMigration < ActiveRecord::Migration[6.0]
  DOWNTIME = true
  DOWNTIME_REASON = 'This migration requires downtime because ...'

  def change
    ...
  end
end

It is an error (that is, CI will fail) if the DOWNTIME constant is missing from a migration class.

Reversibility

Your migration must be reversible. This is very important, as it should be possible to downgrade in case of a vulnerability or bugs.

In your migration, add a comment describing how the reversibility of the migration was tested.

Some migrations cannot be reversed. For example, some data migrations can't be reversed because we lose information about the state of the database before the migration. You should still create a down method with a comment, explaining why the changes performed by the up method can't be reversed, so that the migration itself can be reversed, even if the changes performed during the migration can't be reversed:

def down
  # no-op

  # comment explaining why changes performed by `up` cannot be reversed.
end

Atomicity

By default, migrations are single transaction. That is, a transaction is opened at the beginning of the migration, and committed after all steps are processed.

Running migrations in a single transaction makes sure that if one of the steps fails, none of the steps will be executed, leaving the database in valid state. Therefore, either:

• Put all migrations in one single-transaction migration.
• If necessary, put most actions in one migration and create a separate migration for the steps that cannot be done in a single transaction.

For example, if you create an empty table and need to build an index for it, it is recommended to use a regular single-transaction migration and the default rails schema statement: add_index. This is a blocking operation, but it won't cause problems because the table is not yet used, and therefore it does not have any records yet.

Heavy operations in a single transaction

When using a single-transaction migration, a transaction will hold on a database connection for the duration of the migration, so you must make sure the actions in the migration do not take too much time: In general, queries executed in a migration need to fit comfortably within 15s on GitLab.com.

In case you need to insert, update, or delete a significant amount of data, you:

• Must disable the single transaction with disable_ddl_transaction!.
• Should consider doing it in a Background Migration.

Retry mechanism when acquiring database locks

When changing the database schema, we use helper methods to invoke DDL (Data Definition Language) statements. In some cases, these DDL statements require a specific database lock.

Example:

def change
  remove_column :users, :full_name, :string
end

Executing this migration requires an exclusive lock on the users table. When the table is concurrently accessed and modified by other processes, acquiring the lock may take a while. The lock request is waiting in a queue and it may also block other queries on the users table once it has been enqueued.

More information about PostgresSQL locks: Explicit Locking

For stability reasons, GitLab.com has a specific statement_timeout set. When the migration is invoked, any database query will have a fixed time to execute. In a worst-case scenario, the request will sit in the lock queue, blocking other queries for the duration of the configured statement timeout, then failing with canceling statement due to statement timeout error.

This problem could cause failed application upgrade processes and even application stability issues, since the table may be inaccessible for a short period of time.

To increase the reliability and stability of database migrations, the GitLab codebase offers a helper method to retry the operations with different lock_timeout settings and wait time between the attempts. Multiple smaller attempts to acquire the necessary lock allow the database to process other statements.

Examples

Removing a column:

include Gitlab::Database::MigrationHelpers

def up
  with_lock_retries do
    remove_column :users, :full_name
  end
end

def down
  with_lock_retries do
    add_column :users, :full_name, :string
  end
end

Removing a foreign key:

include Gitlab::Database::MigrationHelpers

def up
  with_lock_retries do
    remove_foreign_key :issues, :projects
  end
end

def down
  with_lock_retries do
    add_foreign_key :issues, :projects
  end
end

Changing default value for a column:

include Gitlab::Database::MigrationHelpers

def up
  with_lock_retries do
    change_column_default :merge_requests, :lock_version, from: nil, to: 0
  end
end

def down
  with_lock_retries do
    change_column_default :merge_requests, :lock_version, from: 0, to: nil
  end
end

Creating a new table with a foreign key:

We can simply wrap the create_table method with with_lock_retries:

def up
  with_lock_retries do
    create_table :issues do |t|
      t.references :project, index: true, null: false, foreign_key: { on_delete: :cascade }
      t.string :title, limit: 255
    end
  end
end

def down
  drop_table :issues
end

Creating a new table when we have two foreign keys:

For this, we'll need three migrations:

1. Creating the table without foreign keys (with the indices).
2. Add foreign key to the first table.
3. Add foreign key to the second table.

Creating the table:

def up
  create_table :imports do |t|
    t.bigint :project_id, null: false
    t.bigint :user_id, null: false
    t.string :jid, limit: 255
  end

  add_index :imports, :project_id
  add_index :imports, :user_id
end

def down
  drop_table :imports
end

Adding foreign key to projects:

include Gitlab::Database::MigrationHelpers

def up
  with_lock_retries do
    add_foreign_key :imports, :projects, column: :project_id, on_delete: :cascade
  end
end

def down
  with_lock_retries do
    remove_foreign_key :imports, column: :project_id
  end
end

Adding foreign key to users:

include Gitlab::Database::MigrationHelpers

def up
  with_lock_retries do
    add_foreign_key :imports, :users, column: :user_id, on_delete: :cascade
  end
end

def down
  with_lock_retries do
    remove_foreign_key :imports, column: :user_id
  end
end

Usage with disable_ddl_transaction!

Generally the with_lock_retries helper should work with disabled_ddl_transaction!. A custom RuboCop rule ensures that only allowed methods can be placed within the lock retries block.

disable_ddl_transaction!

def up
  with_lock_retries do
    add_column :users, :name, :text
  end

  add_text_limit :users, :name, 255 # Includes constraint validation (full table scan)
end

The RuboCop rule generally allows standard Rails migration methods, listed below. This example will cause a Rubocop offense:

disabled_ddl_transaction!

def up
  with_lock_retries do
    add_concurrent_index :users, :name
  end
end

When to use the helper method

The with_lock_retries helper method can be used when you normally use standard Rails migration helper methods. Calling more than one migration helper is not a problem if they're executed on the same table.

Using the with_lock_retries helper method is advised when a database migration involves one of the high-traffic tables:

• users
• projects
• namespaces
• issues
• merge_requests
• ci_pipelines
• ci_builds
• notes

Example changes:

• add_foreign_key / remove_foreign_key
• add_column / remove_column
• change_column_default
• create_table / drop_table

Note: with_lock_retries method cannot be used within the change method, you must manually define the up and down methods to make the migration reversible.

How the helper method works

1. Iterate 50 times.
2. For each iteration, set a pre-configured lock_timeout.
3. Try to execute the given block. (remove_column).
4. If LockWaitTimeout error is raised, sleep for the pre-configured sleep_time and retry the block.
5. If no error is raised, the current iteration has successfully executed the block.

For more information check the Gitlab::Database::WithLockRetries class. The with_lock_retries helper method is implemented in the Gitlab::Database::MigrationHelpers module.

In a worst-case scenario, the method:

• Executes the block for a maximum of 50 times over 40 minutes.
  • Most of the time is spent in a pre-configured sleep period after each iteration.
• After the 50th retry, the block will be executed without lock_timeout, just like a standard migration invocation.
• If a lock cannot be acquired, the migration will fail with statement timeout error.

The migration might fail if there is a very long running transaction (40+ minutes) accessing the users table.

Multi-Threading

Sometimes a migration might need to use multiple Ruby threads to speed up a migration. For this to work your migration needs to include the module Gitlab::Database::MultiThreadedMigration:

class MyMigration < ActiveRecord::Migration[6.0]
  include Gitlab::Database::MigrationHelpers
  include Gitlab::Database::MultiThreadedMigration
end

You can then use the method with_multiple_threads to perform work in separate threads. For example:

class MyMigration < ActiveRecord::Migration[6.0]
  include Gitlab::Database::MigrationHelpers
  include Gitlab::Database::MultiThreadedMigration

  def up
    with_multiple_threads(4) do
      disable_statement_timeout

      # ...
    end
  end
end

Here the call to disable_statement_timeout will use the connection local to the with_multiple_threads block, instead of re-using the global connection pool. This ensures each thread has its own connection object, and won't time out when trying to obtain one.

NOTE: PostgreSQL has a maximum amount of connections that it allows. This limit can vary from installation to installation. As a result, it's recommended you do not use more than 32 threads in a single migration. Usually, 4-8 threads should be more than enough.

Removing indexes

If the table is not empty when removing an index, make sure to use the method remove_concurrent_index instead of the regular remove_index method. The remove_concurrent_index method drops indexes concurrently, so no locking is required, and there is no need for downtime. To use this method, you must disable single-transaction mode by calling the method disable_ddl_transaction! in the body of your migration class like so:

class MyMigration < ActiveRecord::Migration[6.0]
  include Gitlab::Database::MigrationHelpers
  disable_ddl_transaction!

  def up
    remove_concurrent_index :table_name, :column_name
  end
end

Note that it is not necessary to check if the index exists prior to removing it.

For a small table (such as an empty one or one with less than 1,000 records), it is recommended to use remove_index in a single-transaction migration, combining it with other operations that don't require disable_ddl_transaction!.

Adding indexes

Before adding an index, consider if this one is necessary. There are situations in which an index might not be required, like:

• The table is small (less than 1,000 records) and it's not expected to exponentially grow in size.
• Any existing indexes filter out enough rows.
• The reduction in query timings after the index is added is not significant.

Additionally, wide indexes are not required to match all filter criteria of queries, we just need to cover enough columns so that the index lookup has a small enough selectivity. Please review our Adding Database indexes guide for more details.

When adding an index to a non-empty table make sure to use the method add_concurrent_index instead of the regular add_index method. The add_concurrent_index method automatically creates concurrent indexes when using PostgreSQL, removing the need for downtime.

To use this method, you must disable single-transactions mode by calling the method disable_ddl_transaction! in the body of your migration class like so:

class MyMigration < ActiveRecord::Migration[6.0]
  include Gitlab::Database::MigrationHelpers

  DOWNTIME = false

  disable_ddl_transaction!

  def up
    add_concurrent_index :table, :column
  end

  def down
    remove_concurrent_index :table, :column
  end
end

If you need to add a unique index, please keep in mind there is the possibility of existing duplicates being present in the database. This means that should always first add a migration that removes any duplicates, before adding the unique index.

For a small table (such as an empty one or one with less than 1,000 records), it is recommended to use add_index in a single-transaction migration, combining it with other operations that don't require disable_ddl_transaction!.

Adding foreign-key constraints

When adding a foreign-key constraint to either an existing or a new column also remember to add an index on the column.

This is required for all foreign-keys, e.g., to support efficient cascading deleting: when a lot of rows in a table get deleted, the referenced records need to be deleted too. The database has to look for corresponding records in the referenced table. Without an index, this will result in a sequential scan on the table, which can take a long time.

Here's an example where we add a new column with a foreign key constraint. Note it includes index: true to create an index for it.

class Migration < ActiveRecord::Migration[6.0]

  def change
    add_reference :model, :other_model, index: true, foreign_key: { on_delete: :cascade }
  end
end

When adding a foreign-key constraint to an existing column in a non-empty table, we have to employ add_concurrent_foreign_key and add_concurrent_index instead of add_reference.

For an empty table (such as a fresh one), it is recommended to use add_reference in a single-transaction migration, combining it with other operations that don't require disable_ddl_transaction!.

You can read more about adding foreign key constraints to an existing column.

NOT NULL constraints

Introduced in GitLab 13.0.

See the style guide on NOT NULL constraints for more information.

Adding Columns With Default Values

With PostgreSQL 11 being the minimum version since GitLab 13.0, adding columns with default values has become much easier and the standard add_column helper should be used in all cases.

Before PostgreSQL 11, adding a column with a default was problematic as it would have caused a full table rewrite. The corresponding helper add_column_with_default has been deprecated and will be removed in a later release.

NOTE: Note: If a backport adding a column with a default value is needed for %12.9 or earlier versions, it should use add_column_with_default helper. If a large table is involved, backporting to %12.9 is contraindicated.

Changing the column default

One might think that changing a default column with change_column_default is an expensive and disruptive operation for larger tables, but in reality it's not.

Take the following migration as an example:

class DefaultRequestAccessGroups < ActiveRecord::Migration[5.2]
  DOWNTIME = false

  def change
    change_column_default(:namespaces, :request_access_enabled, from: false, to: true)
  end
end

Migration above changes the default column value of one of our largest tables: namespaces. This can be translated to:

ALTER TABLE namespaces
ALTER COLUMN request_access_enabled
DEFAULT false

In this particular case, the default value exists and we're just changing the metadata for request_access_enabled column, which does not imply a rewrite of all the existing records in the namespaces table. Only when creating a new column with a default, all the records are going be rewritten.

NOTE: Note: A faster ALTER TABLE ADD COLUMN with a non-null default was introduced on PostgresSQL 11.0, removing the need of rewriting the table when a new column with a default value is added.

For the reasons mentioned above, it's safe to use change_column_default in a single-transaction migration without requiring disable_ddl_transaction!.

Updating an existing column

To update an existing column to a particular value, you can use update_column_in_batches. This will split the updates into batches, so we don't update too many rows at in a single statement.

This updates the column foo in the projects table to 10, where some_column is 'hello':

update_column_in_batches(:projects, :foo, 10) do |table, query|
  query.where(table[:some_column].eq('hello'))
end

If a computed update is needed, the value can be wrapped in Arel.sql, so Arel treats it as an SQL literal. It's also a required deprecation for Rails 6.

The below example is the same as the one above, but the value is set to the product of the bar and baz columns:

update_value = Arel.sql('bar * baz')

update_column_in_batches(:projects, :foo, update_value) do |table, query|
  query.where(table[:some_column].eq('hello'))
end

Like add_column_with_default, there is a RuboCop cop to detect usage of this on large tables. In the case of update_column_in_batches, it may be acceptable to run on a large table, as long as it is only updating a small subset of the rows in the table, but do not ignore that without validating on the GitLab.com staging environment - or asking someone else to do so for you - beforehand.

Dropping a database table

Dropping a database table is uncommon, and the drop_table method provided by Rails is generally considered safe. Before dropping the table, please consider the following:

If your table has foreign keys on a high-traffic table (like projects), then the DROP TABLE statement might fail with statement timeout error. Determining what tables are high traffic can be difficult. Self-managed instances might use different features of GitLab with different usage patterns, thus making assumptions based on GitLab.com is not enough.

Table has no records (feature was never in use) and no foreign keys:

• Simply use the drop_table method in your migration.

def change
  drop_table :my_table
end

Table has records but no foreign keys:

• First release: Remove the application code related to the table, such as models, controllers and services.
• Second release: Use the drop_table method in your migration.

def up
  drop_table :my_table
end

def down
  # create_table ...
end

Table has foreign keys:

• First release: Remove the application code related to the table, such as models, controllers, and services.
• Second release: Remove the foreign keys using the with_lock_retries helper method. Use drop_table in another migration file.

Migrations for the second release:

Removing the foreign key on the projects table:

# first migration file

def up
  with_lock_retries do
    remove_foreign_key :my_table, :projects
  end
end

def down
  with_lock_retries do
    add_foreign_key :my_table, :projects
  end
end

Dropping the table:

# second migration file

def up
  drop_table :my_table
end

def down
  # create_table ...
end

Integer column type

By default, an integer column can hold up to a 4-byte (32-bit) number. That is a max value of 2,147,483,647. Be aware of this when creating a column that will hold file sizes in byte units. If you are tracking file size in bytes, this restricts the maximum file size to just over 2GB.

To allow an integer column to hold up to an 8-byte (64-bit) number, explicitly set the limit to 8-bytes. This will allow the column to hold a value up to 9,223,372,036,854,775,807.

Rails migration example:

add_column(:projects, :foo, :integer, default: 10, limit: 8)

Strings and the Text data type

Introduced in GitLab 13.0.

See the text data type style guide for more information.

Timestamp column type

By default, Rails uses the timestamp data type that stores timestamp data without timezone information. The timestamp data type is used by calling either the add_timestamps or the timestamps method.

Also, Rails converts the :datetime data type to the timestamp one.

Example:

# timestamps
create_table :users do |t|
  t.timestamps
end

# add_timestamps
def up
  add_timestamps :users
end

# :datetime
def up
  add_column :users, :last_sign_in, :datetime
end

Instead of using these methods, one should use the following methods to store timestamps with timezones:

• add_timestamps_with_timezone
• timestamps_with_timezone
• datetime_with_timezone

This ensures all timestamps have a time zone specified. This, in turn, means existing timestamps won't suddenly use a different timezone when the system's timezone changes. It also makes it very clear which timezone was used in the first place.

Storing JSON in database

The Rails 5 natively supports JSONB (binary JSON) column type. Example migration adding this column:

class AddOptionsToBuildMetadata < ActiveRecord::Migration[5.0]
  DOWNTIME = false

  def change
    add_column :ci_builds_metadata, :config_options, :jsonb
  end
end

You have to use a serializer to provide a translation layer:

class BuildMetadata
  serialize :config_options, Serializers::JSON # rubocop:disable Cop/ActiveRecordSerialize
end

When using a JSONB column, use the JsonSchemaValidator to keep control of the data being inserted over time.

class BuildMetadata
  validates :config_options, json_schema: { filename: 'build_metadata_config_option' }
end

Testing

See the Testing Rails migrations style guide.

Data migration

Please prefer Arel and plain SQL over usual ActiveRecord syntax. In case of using plain SQL, you need to quote all input manually with quote_string helper.

Example with Arel:

users = Arel::Table.new(:users)
users.group(users[:user_id]).having(users[:id].count.gt(5))

#update other tables with these results

Example with plain SQL and quote_string helper:

select_all("SELECT name, COUNT(id) as cnt FROM tags GROUP BY name HAVING COUNT(id) > 1").each do |tag|
  tag_name = quote_string(tag["name"])
  duplicate_ids = select_all("SELECT id FROM tags WHERE name = '#{tag_name}'").map{|tag| tag["id"]}
  origin_tag_id = duplicate_ids.first
  duplicate_ids.delete origin_tag_id

  execute("UPDATE taggings SET tag_id = #{origin_tag_id} WHERE tag_id IN(#{duplicate_ids.join(",")})")
  execute("DELETE FROM tags WHERE id IN(#{duplicate_ids.join(",")})")
end

If you need more complex logic, you can define and use models local to a migration. For example:

class MyMigration < ActiveRecord::Migration[6.0]
  class Project < ActiveRecord::Base
    self.table_name = 'projects'
  end

  def up
    # Reset the column information of all the models that update the database
    # to ensure the Active Record's knowledge of the table structure is current
    Project.reset_column_information

    # ... ...
  end
end

When doing so be sure to explicitly set the model's table name, so it's not derived from the class name or namespace.

Finally, make sure that reset_column_information is run in the up method of the migration for all local Models that update the database.

The reason for that is that all migration classes are loaded at the beginning (when db:migrate starts), so they can get out of sync with the table schema they map to in case another migration updates that schema. That makes the data migration fail when trying to insert or make updates to the underlying table, as the new columns are reported as unknown attribute by ActiveRecord.

Renaming reserved paths

When a new route for projects is introduced, it could conflict with any existing records. The path for these records should be renamed, and the related data should be moved on disk.

Since we had to do this a few times already, there are now some helpers to help with this.

To use this you can include Gitlab::Database::RenameReservedPathsMigration::V1 in your migration. This will provide 3 methods which you can pass one or more paths that need to be rejected.

rename_root_paths: This will rename the path of all namespaces with the given name that don't have a parent_id.

rename_child_paths: This will rename the path of all namespaces with the given name that have a parent_id.

rename_wildcard_paths: This will rename the path of all projects, and all namespaces that have a project_id.

The path column for these rows will be renamed to their previous value followed by an integer. For example: users would turn into users0