When we want to purge all deleted rows we use "glance-manage db purge" or "glance-manage purge_images_table" then we never know how many deleted rows are still in a table. So we need to launch the command many times until the command reports that 0 rows were deleted in every table. It's inconvenient. The patch introduced new valid value for "max_rows". If "max_rows" equals -1 all deleted rows are purged. Closes-Bug: 1702445 Change-Id: Ia9811d93b1c0b67a1b29a9e8ee3e0fc098c57d4a
9.0 KiB
Database Management
Updating and Migrating the Database
The default metadata driver for Glance uses SQLAlchemy, which implies there
exists a backend database which must be managed. The
glance-manage
binary provides a set of commands for making
this easier.
The commands should be executed as a subcommand of 'db':
glance-manage db <cmd> <args>
Note
In the Ocata release (14.0.0), the database migration engine was
changed from SQLAlchemy Migrate to Alembic. This
necessitated some changes in the glance-manage
tool. While
the user interface has been kept as similar as possible, the
glance-manage
tool included with the Ocata and more recent
releases is incompatible with the "legacy" tool. If you are consulting
these documents for information about the glance-manage
tool in the Newton or earlier releases, please see the legacy-database-management
page.
Migration Scripts
The migration scripts are stored in the directory:
glance/db/sqlalchemy/alembic_migrations/versions
As mentioned above, these scripts utilize the Alembic migration
engine, which was first introduced in the Ocata release. All database
migrations up through the Liberty release are consolidated into one
Alembic migration script named liberty_initial
. Mitaka
migrations are retained, but have been rewritten for Alembic and named
using the new naming convention.
A fresh Glance installation will apply the following migrations:
liberty-initial
mitaka01
mitaka02
ocata01
Note
The "old-style" migration scripts have been retained in their current directory in the Ocata release so that interested operators can correlate them with the new migrations. This directory will be removed in future releases.
In particular, the "old-style" script for the Ocata migration, 045_add_visibility.py is retained for operators who are conversant in SQLAlchemy Migrate and are interested in comparing it with a "new-style" Alembic migration script. The Alembic script, which is the one actually used to do the upgrade to Ocata, is ocata01_add_visibility_remove_is_public.py.
Sync the Database
glance-manage db sync [VERSION]
Place an existing database under migration control and upgrade it to the specified VERSION or to the latest migration level if VERSION is not specified.
Note
Prior to Ocata release the database version was a numeric value. For
example: for the Newton release, the latest migration level was
44
. Starting with Ocata, database version is a revision
name corresponding to the latest migration included in the release. For
the Ocata release, there is only one database migration and it is
identified by revision ocata01
. So, the database version
for Ocata release is ocata01
.
This naming convention will change slightly with the introduction of
zero-downtime upgrades, which is EXPERIMENTAL in Ocata, but is projected
to be the official upgrade method beginning with the Pike release. See
zero-downtime
for more
information.
Determining the Database Version
glance-manage db version
This will print the current migration level of a Glance database.
Upgrading an Existing Database
glance-manage db upgrade [VERSION]
This will take an existing database and upgrade it to the specified VERSION.
Downgrading an Existing Database
Downgrading an existing database is NOT SUPPORTED.
Upgrades involve complex operations and can fail. Before attempting any upgrade, you should make a full database backup of your production data. As of the OpenStack Kilo release (April 2013), database downgrades are not supported, and the only method available to get back to a prior database version is to restore from backup.
Database Maintenance
Like most OpenStack systems, Glance performs soft deletions
when it deletes records from its database. Depending on usage patterns
in your cloud, you may occasionally want to actually remove such soft
deleted table rows. This operation is called purging the
database, and you can use the glance-manage
tool to do
this.
High-Level Database Architecture
Roughly, what we've got in the glance database is an images table that stores the image id and some other core image properties. All the other information about the image (for example: where the image data is stored in the backend, what projects an image has been shared with, image tags, custom image properties) is stored in other tables in which the image id is a foreign key.
Because the images table keeps track of what image identifiers have been issued, it must be treated differently from the other tables with respect to purging the database.
Note
Before the Rocky release (17.0.0), the images table was not treated differently, which made Glance vulnerable to OSSN-0075, "Deleted Glance image IDs may be reassigned". Please read through that OpenStack Security Note to understand the nature of the problem.
Additionally, the Glance spec Mitigate
OSSN-0075 contains a discussion of the issue and explains the
changes made to the glance-manage
tool for the Rocky
release. The Gerrit
review of the spec contains an extensive discussion of several
alternative approaches and will give you an idea of why the Glance team
provided a "mitigation" instead of a fix.
Purging the Database
You can use the glance-manage
tool to purge the
soft-deleted rows from all tables except the images table:
glance-manage db purge
This command takes two optional parameters:
- --age_in_days NUM Only purge rows that have been deleted for longer
-
than NUM days. The default is 30 days.
- --max_rows NUM Purge a maximum of NUM rows from each table.
-
The default is 100. All deleted rows are purged if equals -1
Purging the Images Table
Remember that image identifiers are used by other OpenStack services that require access to images. These services expect that when an image is requested by ID, they will receive the same data every time. When the images table is purged of its soft-deleted rows, Glance loses its memory that those image IDs were ever mapped to some particular payload. Thus, care must be taken in purging the images table. We recommend that it be done much less frequently than the "regular" purge operation.
Use the following command to purge the images table:
glance-manage db purge_images_table
Be sure you have read and understood the implications of OSSN-0075 before you use this command, which purges the soft-deleted rows from the images table.
It takes two optional parameters:
- --age_in_days NUM Only purge rows that have been deleted for longer
-
than NUM days. The default is 180 days.
- --max_rows NUM Purge a maximum of NUM rows from the images table.
-
The default is 100. All deleted rows are purged if equals -1
It is possible for this command to fail with an IntegrityError saying
something like "Cannot delete or update a parent row: a foreign key
constraint fails". This can happen when you try to purge records from
the images table when related records have not yet been
purged from other tables. The purge_images_table
command
should only be issued after all related information has been purged
using the "regular" purge
command.