deb-glance/doc/source/configuring.rst

Configuring Glance

Glance has a number of options that you can use to configure the Glance API server, the Glance Registry server, and the various storage backends that Glance can use to store images.

Most configuration is done via configuration files, with the Glance API server and Glance Registry server using separate configuration files.

When starting up a Glance server, you can specify the configuration file to use (see the documentation on controller Glance servers). If you do not specify a configuration file, Glance will look in the following directories for a configuration file, in order:

• ~/.glance
• ~/
• /etc/glance
• /etc

The Glance API server configuration file should be named glance-api.conf. Similarly, the Glance Registry server configuration file should be named glance-registry.conf. If you installed Glance via your operating system's package management system, it is likely that you will have sample configuration files installed in /etc/glance.

In addition to this documentation page, you can check the etc/glance-api.conf and etc/glance-registry.conf sample configuration files distributed with Glance for example configuration files for each server application with detailed comments on what each options does.

The PasteDeploy configuration (controlling the deployment of the WSGI application for each component) may be found by default in <component>-paste.ini alongside the main configuration file, <component>.conf. For example, glance-api-paste.ini corresponds to glance-api.conf. This pathname for the paste config is configurable, as follows:

[paste_deploy]
config_file = /path/to/paste/config

Common Configuration Options in Glance

Glance has a few command-line options that are common to all Glance programs:

• --verbose

Optional. Default: False

Can be specified on the command line and in configuration files.

Turns on the INFO level in logging and prints more verbose command-line interface printouts.

• --debug

Optional. Default: False

Can be specified on the command line and in configuration files.

Turns on the DEBUG level in logging.

• --config-file=PATH

Optional. Default: See below for default search order.

Specified on the command line only.

Takes a path to a configuration file to use when running the program. If this CLI option is not specified, then we check to see if the first argument is a file. If it is, then we try to use that as the configuration file. If there is no file or there were no arguments, we search for a configuration file in the following order:

• ~/.glance
• ~/
• /etc/glance
• /etc

The filename that is searched for depends on the server application name. So, if you are starting up the API server, glance-api.conf is searched for, otherwise glance-registry.conf.

• --config-dir=DIR

Optional. Default: None

Specified on the command line only.

Takes a path to a configuration directory from which all *.conf fragments are loaded. This provides an alternative to multiple --config-file options when it is inconvenient to explicitly enumerate all the config files, for example when an unknown number of config fragments are being generated by a deployment framework.

If --config-dir is set, then --config-file is ignored.

An example usage would be:

$ glance-api --config-dir=/etc/glance/glance-api.d

$ ls /etc/glance/glance-api.d

00-core.conf 01-s3.conf 02-swift.conf 03-ssl.conf ... etc.

The numeric prefixes in the example above are only necessary if a specific parse ordering is required (i.e. if an individual config option set in an earlier fragment is overridden in a later fragment).

Configuring Server Startup Options

You can put the following options in the glance-api.conf and glance-registry.conf files, under the [DEFAULT] section. They enable startup and binding behaviour for the API and registry servers, respectively.

• bind_host=ADDRESS

The address of the host to bind to.

Optional. Default: 0.0.0.0

• bind_port=PORT

The port the server should bind to.

Optional. Default: 9191 for the registry server, 9292 for the API server

• backlog=REQUESTS

Number of backlog requests to configure the socket with.

Optional. Default: 4096

• workers=PROCESSES

Number of Glance API worker processes to start. Each worker process will listen on the same port. Increasing this value may increase performance (especially if using SSL with compression enabled). Typically it is recommended to have one worker process per CPU. The value 0 will prevent any new processes from being created.

Optional. Default: 0

Configurating SSL Support

• cert_file=PATH

Path to the certificate file the server should use when binding to an SSL-wrapped socket.

Optional. Default: not enabled.

• key_file=PATH

Path to the private key file the server should use when binding to an SSL-wrapped socket.

Optional. Default: not enabled.

• registry_client_protocol=PROTOCOL

If you run a secure Registry server, you need to set this value to https and also set registry_client_key_file and optionally registry_client_cert_file.

Optional. Default: http

• registry_client_key_file=PATH

The path to the key file to use in SSL connections to the registry server, if any. Alternately, you may set the GLANCE_CLIENT_KEY_FILE environ variable to a filepath of the key file

Optional. Default: Not set.

• registry_client_cert_file=PATH

Optional. Default: Not set.

The path to the cert file to use in SSL connections to the registry server, if any. Alternately, you may set the GLANCE_CLIENT_CERT_FILE environ variable to a filepath of the cert file

• registry_client_ca_file=PATH

Optional. Default: Not set.

The path to a Certifying Authority's cert file to use in SSL connections to the registry server, if any. Alternately, you may set the GLANCE_CLIENT_CA_FILE environ variable to a filepath of the CA cert file

Configuring Logging in Glance

There are a number of configuration options in Glance that control how Glance servers log messages.

• --log-config=PATH

Optional. Default: None

Specified on the command line only.

Takes a path to a configuration file to use for configuring logging.

Logging Options Available Only in Configuration Files

You will want to place the different logging options in the [DEFAULT] section in your application configuration file. As an example, you might do the following for the API server, in a configuration file called etc/glance-api.conf:

[DEFAULT]
log_file = /var/log/glance/api.log

• log_file

The filepath of the file to use for logging messages from Glance's servers. If missing, the default is to output messages to stdout, so if you are running Glance servers in a daemon mode (using glance-control) you should make sure that the log_file option is set appropriately.

• log_dir

The filepath of the directory to use for log files. If not specified (the default) the log_file is used as an absolute filepath.

• log_date_format

The format string for timestamps in the log output.

Defaults to %Y-%m-%d %H:%M:%S. See the logging module documentation for more information on setting this format string.

• log_use_syslog

Use syslog logging functionality.

Defaults to False.

Configuring Glance Storage Backends

There are a number of configuration options in Glance that control how Glance stores disk images. These configuration options are specified in the glance-api.conf config file in the section [DEFAULT].

• default_store=STORE

Optional. Default: file

Can only be specified in configuration files.

Sets the storage backend to use by default when storing images in Glance. Available options for this option are (file, swift, s3, or rbd).

Configuring the Filesystem Storage Backend

• filesystem_store_datadir=PATH

Optional. Default: /var/lib/glance/images/

Can only be specified in configuration files.

This option is specific to the filesystem storage backend.

Sets the path where the filesystem storage backend write disk images. Note that the filesystem storage backend will attempt to create this directory if it does not exist. Ensure that the user that glance-api runs under has write permissions to this directory.

Configuring the Swift Storage Backend

• swift_store_auth_address=URL

Required when using the Swift storage backend.

Can only be specified in configuration files.

This option is specific to the Swift storage backend.

Sets the authentication URL supplied to Swift when making calls to its storage system. For more information about the Swift authentication system, please see the Swift auth documentation and the overview of Swift authentication.

IMPORTANT NOTE: Swift authentication addresses use HTTPS by default. This means that if you are running Swift with authentication over HTTP, you need to set your swift_store_auth_address to the full URL, including the http://.

• swift_store_user=USER

Required when using the Swift storage backend.

Can only be specified in configuration files.

This option is specific to the Swift storage backend.

Sets the user to authenticate against the swift_store_auth_address with.

• swift_store_key=KEY

Required when using the Swift storage backend.

Can only be specified in configuration files.

This option is specific to the Swift storage backend.

Sets the authentication key to authenticate against the swift_store_auth_address with for the user swift_store_user.

• swift_store_container=CONTAINER

Optional. Default: glance

Can only be specified in configuration files.

This option is specific to the Swift storage backend.

Sets the name of the container to use for Glance images in Swift.

• swift_store_create_container_on_put

Optional. Default: False

Can only be specified in configuration files.

This option is specific to the Swift storage backend.

If true, Glance will attempt to create the container swift_store_container if it does not exist.

• swift_store_large_object_size=SIZE_IN_MB

Optional. Default: 5120

Can only be specified in configuration files.

This option is specific to the Swift storage backend.

What size, in MB, should Glance start chunking image files and do a large object manifest in Swift? By default, this is the maximum object size in Swift, which is 5GB

• swift_store_large_object_chunk_size=SIZE_IN_MB

Optional. Default: 200

Can only be specified in configuration files.

This option is specific to the Swift storage backend.

When doing a large object manifest, what size, in MB, should Glance write chunks to Swift? The default is 200MB.

Configuring the S3 Storage Backend

• s3_store_host=URL

Required when using the S3 storage backend.

Can only be specified in configuration files.

This option is specific to the S3 storage backend.

Default: s3.amazonaws.com

Sets the main service URL supplied to S3 when making calls to its storage system. For more information about the S3 authentication system, please see the S3 documentation

• s3_store_access_key=ACCESS_KEY

Required when using the S3 storage backend.

Can only be specified in configuration files.

This option is specific to the S3 storage backend.

Sets the access key to authenticate against the s3_store_host with.

You should set this to your 20-character Amazon AWS access key.

• s3_store_secret_key=SECRET_KEY

Required when using the S3 storage backend.

Can only be specified in configuration files.

This option is specific to the S3 storage backend.

Sets the secret key to authenticate against the s3_store_host with for the access key s3_store_access_key.

You should set this to your 40-character Amazon AWS secret key.

• s3_store_bucket=BUCKET

Required when using the S3 storage backend.

Can only be specified in configuration files.

This option is specific to the S3 storage backend.

Sets the name of the bucket to use for Glance images in S3.

Note that the namespace for S3 buckets is global, therefore you must use a name for the bucket that is unique. It is recommended that you use a combination of your AWS access key, lowercased with "glance".

For instance if your Amazon AWS access key is:

ABCDEFGHIJKLMNOPQRST

then make your bucket value be:

abcdefghijklmnopqrstglance

• s3_store_create_bucket_on_put

Optional. Default: False

Can only be specified in configuration files.

This option is specific to the S3 storage backend.

If true, Glance will attempt to create the bucket s3_store_bucket if it does not exist.

• s3_store_object_buffer_dir=PATH

Optional. Default: the platform's default temporary directory

Can only be specified in configuration files.

This option is specific to the S3 storage backend.

When sending images to S3, what directory should be used to buffer the chunks? By default the platform's temporary directory will be used.

Configuring the RBD Storage Backend

Note: the RBD storage backend requires the python bindings for librados and librbd. These are in the python-ceph package on Debian-based distributions.

• rbd_store_pool=POOL

Optional. Default: rbd

Can only be specified in configuration files.

This option is specific to the RBD storage backend.

Sets the RADOS pool in which images are stored.

• rbd_store_chunk_size=CHUNK_SIZE_MB

Optional. Default: 4

Can only be specified in configuration files.

This option is specific to the RBD storage backend.

Images will be chunked into objects of this size (in megabytes). For best performance, this should be a power of two.

• rbd_store_ceph_conf=PATH

Optional. Default: /etc/ceph/ceph.conf, ~/.ceph/config, and ./ceph.conf

Can only be specified in configuration files.

This option is specific to the RBD storage backend.

Sets the Ceph configuration file to use.

• rbd_store_user=NAME

Optional. Default: admin

Can only be specified in configuration files.

This option is specific to the RBD storage backend.

Sets the RADOS user to authenticate as. This is only needed when RADOS authentication is enabled.

A keyring must be set for this user in the Ceph configuration file, e.g. with a user glance:

[client.glance]
keyring=/etc/glance/rbd.keyring

To set up a user named glance with minimal permissions, using a pool called images, run:

rados mkpool images
ceph-authtool --create-keyring /etc/glance/rbd.keyring
ceph-authtool --gen-key --name client.glance --cap mon 'allow r' --cap osd 'allow rwx pool=images' /etc/glance/rbd.keyring
ceph auth add client.glance -i /etc/glance/rbd.keyring

Configuring the Image Cache

Glance API servers can be configured to have a local image cache. Caching of image files is transparent and happens using a piece of middleware that can optionally be placed in the server application pipeline.

This pipeline is configured in the PasteDeploy configuration file, <component>-paste.ini. You should not generally have to edit this file directly, as it ships with ready-made pipelines for all common deployment flavors.

Enabling the Image Cache Middleware

To enable the image cache middleware, the cache middleware must occur in the application pipeline after the appropriate context middleware.

The cache middleware should be in your glance-api-paste.ini in a section titled [filter:cache]. It should look like this:

[filter:cache]
paste.filter_factory = glance.api.middleware.cache:CacheFilter.factory

A ready-made application pipeline including this filter is defined in the glance-api-paste.ini file, looking like so:

[pipeline:glance-api-caching]
pipeline = versionnegotiation context cache apiv1app

To enable the above application pipeline, in your main glance-api.conf configuration file, select the appropriate deployment flavor like so:

[paste_deploy]
flavor = caching

And that would give you a transparent image cache on the API server.

Configuration Options Affecting the Image Cache

One main configuration file option affects the image cache.

• image_cache_dir=PATH

Required when image cache middleware is enabled.

Default: /var/lib/glance/image-cache

This is the base directory the image cache can write files to. Make sure the directory is writeable by the user running the glance-api server

• image_cache_driver=DRIVER

Optional. Choice of sqlite or xattr

Default: sqlite

The default sqlite cache driver has no special dependencies, other than the python-sqlite3 library, which is installed on virtually all operating systems with modern versions of Python. It stores information about the cached files in a SQLite database.

The xattr cache driver required the python-xattr>=0.6.0 library and requires that the filesystem containing image_cache_dir have access times tracked for all files (in other words, the noatime option CANNOT be set for that filesystem). In addition, user_xattr must be set on the filesystem's description line in fstab. Because of these requirements, the xattr cache driver is not available on Windows.

• image_cache_sqlite_db=DB_FILE

Optional.

Default: cache.db

When using the sqlite cache driver, you can set the name of the database that will be used to store the cached images information. The database is always contained in the image_cache_dir.

• image_cache_max_size=SIZE

Optional.

Default: 10737418240 (10 GB)

Size, in bytes, that the image cache should be constrained to. Images files are cached automatically in the local image cache, even if the writing of that image file would put the total cache size over this size. The glance-cache-pruner executable is what prunes the image cache to be equal to or less than this value. The glance-cache-pruner executable is designed to be run via cron on a regular basis. See more about this executable in Controlling the Growth of the Image Cache <cache>

Note

These configuration options must be set in both the glance-cache and glance-api configuration files.

Configuring the Glance Registry

There are a number of configuration options in Glance that control how this registry server operates. These configuration options are specified in the glance-registry.conf config file in the section [DEFAULT].

• sql_connection=CONNECTION_STRING (--sql-connection when specified on command line)

Optional. Default: None

Can be specified in configuration files. Can also be specified on the command-line for the glance-manage program.

Sets the SQLAlchemy connection string to use when connecting to the registry database. Please see the documentation for SQLAlchemy connection strings online.

• sql_timeout=SECONDS on command line)

Optional. Default: 3600

Can only be specified in configuration files.

Sets the number of seconds after which SQLAlchemy should reconnect to the datastore if no activity has been made on the connection.

Configuring Notifications

Glance can optionally generate notifications to be logged or sent to a RabbitMQ queue. The configuration options are specified in the glance-api.conf config file in the section [DEFAULT].

• notifier_strategy

Optional. Default: noop

Sets the strategy used for notifications. Options are logging, rabbit, qpid and noop. For more information Glance notifications <notifications>

• rabbit_host

Optional. Default: localhost

Host to connect to when using rabbit strategy.

• rabbit_port

Optional. Default: 5672

Port to connect to when using rabbit strategy.

• rabbit_use_ssl

Optional. Default: false

Boolean to use SSL for connecting when using rabbit strategy.

• rabbit_userid

Optional. Default: guest

Userid to use for connection when using rabbit strategy.

• rabbit_password

Optional. Default: guest

Password to use for connection when using rabbit strategy.

• rabbit_virtual_host

Optional. Default: /

Virtual host to use for connection when using rabbit strategy.

• rabbit_notification_exchange

Optional. Default: glance

Exchange name to use for connection when using rabbit strategy.

• rabbit_notification_topic

Optional. Default: glance_notifications

Topic to use for connection when using rabbit strategy.

• rabbit_max_retries

Optional. Default: 0

Number of retries on communication failures when using rabbit strategy. A value of 0 means to retry forever.

• rabbit_retry_backoff

Optional. Default: 2

Number of seconds to wait before reconnecting on failures when using rabbit strategy.

• rabbit_retry_max_backoff

Optional. Default: 30

Maximum seconds to wait before reconnecting on failures when using rabbit strategy.

• qpid_notification_exchange

Optional. Default: glance

Message exchange to use when using the qpid notification strategy.

• qpid_notification_topic

Optional. Default: glanice_notifications

This is the topic prefix for notifications when using the qpid notification strategy. When a notification is sent at the info priority, the topic will be glance_notifications.info. The same idea applies for the error and warn notification priorities. To receive all notifications, you would set up a receiver with a topic of glance_notifications.*.

• qpid_host

Optional. Default: localhost

This is the hostname or IP address of the Qpid broker that will be used when Glance has been configured to use the qpid notification strategy.

• qpid_port

Optional. Default: 5672

This is the port number to connect to on the Qpid broker, qpid_host, when using the qpid notification strategy.

• qpid_username

Optional. Default: None

This is the username that Glance will use to authenticate with the Qpid broker if using the qpid notification strategy.

• qpid_password

Optional. Default: None

This is the username that Glance will use to authenticate with the Qpid broker if using the qpid notification strategy.

• qpid_sasl_mechanisms

Optional. Default: None

This is a space separated list of SASL mechanisms to use for authentication with the Qpid broker if using the qpid notification strategy.

• qpid_reconnect_timeout

Optional. Default: None

This option specifies a timeout in seconds for automatic reconnect attempts to the Qpid broker if the qpid notification strategy is used. In general, it is safe to leave all of the reconnect timing options not set. In that case, the Qpid client's default behavior will be used, which is to attempt to reconnect to the broker at exponential back-off intervals (in 1 second, then 2 seconds, then 4, 8, 16, etc).

• qpid_reconnect_limit

Optional. Default: None

This option specifies a maximum number of reconnect attempts to the Qpid broker if the qpid notification strategy is being used. Normally the Qpid client will continue attempting to reconnect until successful.

• qpid_reconnect_interval_min

Optional. Default: None

This option specifies the minimum number of seconds between reconnection attempts if the qpid notification strategy is being used.

• qpid_reconnect_interval_max

Optional. Default: None

This option specifies the maximum number of seconds between reconnection attempts if the qpid notification strategy is being used.

• qpid_reconnect_interval

This option specifies the exact number of seconds between reconnection attempts if the qpid notification strategy is being used. Setting this option is equivalent to setting qpid_reconnect_interval_max and qpid_reconnect_interval_min to the same value.

• qpid_heartbeat

Optional. Default: 5

This option is used to specify the number of seconds between heartbeat messages exchanged between the Qpid client and Qpid broker if the qpid notification strategy is being used. Heartbeats are used to more quickly detect that a connection has been lost.

• qpid_protocol

Optional. Default: tcp

This option is used to specify the transport protocol to use if using the qpid notification strategy. To enable SSL, set this option to ssl.

• qpid_tcp_nodelay

Optional. Default: True

This option can be used to disable the TCP NODELAY option. It effectively disables the Nagle algorithm for the connection to the Qpid broker. This option only applies if the qpid notification strategy is used.

Configuring Access Policies

Access rules may be configured using a Policy Configuration file <policies>. Two configuration options tell the Glance API server about the policies to use.

• policy_file=PATH

Optional. Default: Looks for a file called policy.json or glance.policy.json in standard configuration directories.

Policy file to load when starting the API server

• policy_default_rule=RULE

Optional. Default: "default"

Name of the rule in the policy configuration file to use as the default rule