Once you have deployed a Project Quay registry, there are many ways you can further configure and manage that deployment. Topics covered here include:

  • Advanced Project Quay configuration

  • Setting notifications to alert you of a new Project Quay release

  • Securing connections with SSL and TLS certificates

  • Directing action logs storage to Elasticsearch

  • Configuring image security scanning with Clair

  • Scan pod images with the Container Security Operator

  • Integrate Project Quay into OpenShift with the Quay Bridge Operator

  • Mirroring images with repository mirroring

  • Sharing Quay images with a BitTorrent service

  • Authenticating users with LDAP

  • Enabling Quay for Prometheus and Grafana metrics

  • Setting up geo-replication

  • Troubleshooting Quay

Advanced Project Quay configuration

You can configure your Project Quay after initial deployment using several different interfaces:

  • The Project Quay Config Tool: Running the Quay container in config mode presents a Web-based interface for configuring the Project Quay cluster. This is the recommended method for most configuration of the Project Quay service itself.

  • Editing the config.yaml: The config.yaml file holds most of the configuration information for the Project Quay cluster. Editing that file directly is possible, but it is only recommended for advanced tuning and performance features that are not available through the Config Tool.

  • Project Quay API: Some Project Quay configuration can be done through the API.

While configuration for specific features is covered in separate sections, this section describes how to use each of those interfaces and perform some more advanced configuration.

Using Project Quay Config Tool to modify Project Quay

The Project Quay Config Tool is made available by running a Quay container in config mode alongside the regular Project Quay service. Running the Config Tool is different for Project Quay clusters running on OpenShift than it is for those running directly on host systems.

Running the Config Tool from the Project Quay Operator

If you are running the Project Quay Operator from OpenShift, the Config Tool is probably already available for you to use. To access the Config Tool, do the following:

  1. From the OpenShift console, select the project in which Project Quay is running. For example, quay-enterprise.

  2. From the left column, select Networking → Routes. You should see routes to both the Project Quay application and Config Tool, as shown in the following image:

    View the route to the Project Quay Config Tool

  3. Select the route to the Config Tool (for example, example-quayecosystem-quay-config) and select it. The Config tool Web UI should open in your browser.

  4. Select Modify configuration for this cluster. You should see the Config Tool, ready for you to change features of your Project Quay cluster, as shown in the following image:

    Modify Project Quay cluster settings from the Config Tool

  5. When you have made the changes you want, select Save Configuration Changes. The Config Tool will validate your changes.

  6. Make any corrections as needed by selecting Continue Editing or select Next to continue on.

  7. When prompted, it is recommended that you select Download Configuration. That will download a tarball of your new config.yaml, as well as any certificates and keys used with your Project Quay setup.

  8. Select Go to deployment rollout, then Populate the configuration to deployments. The Project Quay pods will be restarted and the changes will take effect.

The config.yaml file you saved can be used to make advanced changes to your configuration or just kept for future reference.

Running the Config Tool from the command line

If you are running Project Quay directly from a host system, using tools such as the podman or docker commands, after the initial Project Quay deployment, you can restart the Config Tool to modify your Project Quay cluster. Here’s how:

  1. Start quay in config mode: On the first quay node run the following, replacing my-secret-password with your password. If you would like to modify an existing config bundle, you can simply mount your configuration directory into the Quay container as you would in registry mode.

    # podman run --rm -it --name quay_config -p 8080:8080 \
        -v path/to/config-bundle:/conf/stack \
        quay.io/projectquay/quay:v3.7.1 config my-secret-password
  2. Open browser: When the quay configuration tool starts up, open a browser to the URL and port 8080 of the system you are running the configuration tool on (for example https://myquay.example.com:8080). You are prompted for a username and password.

At this point, you can begin modifying your Project Quay cluster as described earlier.

Using the API to modify Project Quay

See the Project Quay API Guide for information on how to access Project Quay API.

Editing the config.yaml file to modify Project Quay

Some advanced Project Quay configuration that is not available through the Config Tool can be achieved by editing the config.yaml file directly. Available settings are described in the Schema for Project Quay configuration The following are examples of settings you can change directly in the config.yaml file.

Add name and company to Project Quay sign-in

Setting the following will cause users to be prompted for their name and company when they first sign in. Although this is optional, it can provide you with extra data about your Project Quay users:

+ FEATURE_USER_METADATA: true

Disable TLS Protocols

You can change the SSL_PROTOCOLS setting to remove SSL protocols that you do not want to support in your Project Quay instance. For example, to remove TLS v1 support from the default SSL_PROTOCOLS : ['TLSv1','TLSv1.1','TLSv1.2'], change it as follows:

+ SSL_PROTOCOLS : ['TLSv1.1','TLSv1.2']

Rate limit API calls

Adding the FEATURE_RATE_LIMITS parameter to the config.yaml causes nginx to limit certain API calls to 30 per second. If that feature is not set, API calls are limied to 300 per second (effectively unlimited). Rate limiting can be an important feature, if you need to make sure the resources available are not overwhelmed with traffic.

Some namespace may require unlimited access (perhaps they are important to CI/CD and take priority, for example). In this case, those namespace may be placed in a list in config.yaml for NON_RATE_LIMITED_NAMESPACES.

Adjust database connection pooling

Project Quay is composed of many different processes which all run within the same container. Many of these processes interact with the database.

If enabled, each process that interacts with the database will contain a connection pool. These per-process connection pools are configured to maintain a maximum of 20 connections. Under heavy load, it is possible to fill the connection pool for every process within a Project Quay container. Under certain deployments and loads, this may require analysis to ensure Project Quay does not exceed the database’s configured maximum connection count.

Overtime, the connection pools will release idle connections. To release all connections immediately, Project Quay requires a restart.

Database connection pooling may be toggled by setting the environment variable DB_CONNECTION_POOLING={true|false}

If database connection pooling is enabled, it is possible to change the maximum size of the connection pool. This can be done through the following config.yaml option:

DB_CONNECTION_ARGS:
  max_connections: 10
Database connection arguments

You can customize Project Quay database connection settings within the config.yaml file. These are entirely dependent upon the underlying database driver, such as psycopg2 for Postgres and pymysql for MySQL. It is also possible to pass in arguments used by Peewee’s Connection Pooling mechanism as seen below.

DB_CONNECTION_ARGS:
  max_connections: n  # Max Connection Pool size. (Connection Pooling only)
  timeout: n  # Time to hold on to connections. (Connection Pooling only)
  stale_timeout: n  # Number of seconds to block when the pool is full. (Connection Pooling only)
Database SSL configuration

Some key-value pairs defined under DB_CONNECTION_ARGS are generic while others are database-specific. In particular, SSL configuration depends on the database you are deploying.

PostgreSQL SSL connection arguments

A sample PostgreSQL SSL configuration is given below:

DB_CONNECTION_ARGS:
  sslmode: verify-ca
  sslrootcert: /path/to/cacert

The sslmode option determines whether or with what priority a secure SSL TCP/IP connection will be negotiated with the server. There are six modes:

  • disable: only try a non-SSL connection

  • allow: first try a non-SSL connection; if that fails, try an SSL connection

  • prefer: (default) first try an SSL connection; if that fails, try a non-SSL connection

  • require: only try an SSL connection. If a root CA file is present, verify the certificate in the same way as if verify-ca was specified

  • verify-ca: only try an SSL connection, and verify that the server certificate is issued by a trusted certificate authority (CA)

  • verify-full: only try an SSL connection, verify that the server certificate is issued by a trusted CA and that the requested server host name matches that in the certificate

More information on the valid arguments for PostgreSQL is available at https://www.postgresql.org/docs/current/libpq-connect.html.

MySQL SSL connection arguments

A sample MySQL SSL configuration follows:

DB_CONNECTION_ARGS:
  ssl:
    ca: /path/to/cacert

Information on the valid connection arguments for MySQL is available at https://dev.mysql.com/doc/refman/8.0/en/connecting-using-uri-or-key-value-pairs.html.

HTTP connection counts

It is possible to specify the quantity of simultaneous HTTP connections using environment variables. These can be specified as a whole, or for a specific component. The default for each is 50 parallel connections per process.

Environment variables:

WORKER_CONNECTION_COUNT_REGISTRY=n
WORKER_CONNECTION_COUNT_WEB=n
WORKER_CONNECTION_COUNT_SECSCAN=n
WORKER_CONNECTION_COUNT=n
Note

Specifying a count for a specific component will override any value set in WORKER_CONNECTION_COUNT.

Dynamic process counts

To estimate the quantity of dynamically sized processes, the following calculation is used by default.

Note
Project Quay queries the available CPU count from the entire machine. Any limits applied using kubernetes or other non-virtualized mechanisms will not affect this behavior; Project Quay will makes its calculation based on the total number of processors on the Node. The default values listed are simply targets, but shall not exceed the maximum or be lower than the minimum.

Each of the following process quantities can be overridden using the environment variable specified below.

  • registry - Provides HTTP endpoints to handle registry action

    • minimum: 8

    • maximum: 64

    • default: $CPU_COUNT x 4

    • environment variable: WORKER_COUNT_REGISTRY

  • web - Provides HTTP endpoints for the web-based interface

    • minimum: 2

    • maximum: 32

    • default: $CPU_COUNT x 2

    • environment_variable: WORKER_COUNT_WEB

  • secscan - Interacts with Clair

    • minimum: 2

    • maximum: 4

    • default: $CPU_COUNT x 2

    • environment variable: WORKER_COUNT_SECSCAN

Environment variables

Project Quay allows overriding default behavior using environment variables. This table lists and describes each variable and the values they can expect.

Table 1. Worker count environment variables
Variable Description Values

WORKER_COUNT_REGISTRY

Specifies the number of processes to handle Registry requests within the Quay container.

Integer between 8 and 64

WORKER_COUNT_WEB

Specifies the number of processes to handle UI/Web requests within the container.

Integer between 2 and 32

WORKER_COUNT_SECSCAN

Specifies the number of processes to handle Security Scanning (e.g. Clair) integration within the container.

Integer between 2 and 4

DB_CONNECTION_POOLING

Toggle database connection pooling. In 3.4, it is disabled by default.

"true" or "false"

Turning off connection pooling

Project Quay deployments with a large amount of user activity can regularly hit the 2k maximum database connection limit. In these cases, connection pooling, which is enabled by default for Project Quay, can cause database connection count to rise exponentially and require you to turn off connection pooling.

If turning off connection pooling is not enough to prevent hitting that 2k database connection limit, you need to take additional steps to deal with the problem. In this case you might need to increase the maximum database connections to better suit your workload.

Using the configuration API

The configuration tool exposes 4 endpoints that can be used to build, validate, bundle and deploy a configuration. The config-tool API is documented at https://github.com/quay/config-tool/blob/master/pkg/lib/editor/API.md. In this section, you will see how to use the API to retrieve the current configuration and how to validate any changes you make.

Retrieving the default configuration

If you are running the configuration tool for the first time, and do not have an existing configuration, you can retrieve the default configuration. Start the container in config mode:

$ sudo podman run --rm -it --name quay_config \
  -p 8080:8080 \
  quay.io/projectquay/quay:v3.7.1 config secret

Use the config endpoint of the configuration API to get the default:

$ curl -X GET -u quayconfig:secret http://quay-server:8080/api/v1/config  | jq

The value returned is the default configuration in JSON format:

{
  "config.yaml": {
    "AUTHENTICATION_TYPE": "Database",
    "AVATAR_KIND": "local",
    "DB_CONNECTION_ARGS": {
      "autorollback": true,
      "threadlocals": true
    },
    "DEFAULT_TAG_EXPIRATION": "2w",
    "EXTERNAL_TLS_TERMINATION": false,
    "FEATURE_ACTION_LOG_ROTATION": false,
    "FEATURE_ANONYMOUS_ACCESS": true,
    "FEATURE_APP_SPECIFIC_TOKENS": true,
    ....
  }

}

Retrieving the current configuration

If you have already configured and deployed the Quay registry, stop the container and restart it in configuration mode, loading the existing configuration as a volume:

$ sudo podman run --rm -it --name quay_config \
  -p 8080:8080 \
  -v $QUAY/config:/conf/stack:Z \
  quay.io/projectquay/quay:v3.7.1 config secret

Use the config endpoint of the API to get the current configuration:

$ curl -X GET -u quayconfig:secret http://quay-server:8080/api/v1/config  | jq

The value returned is the current configuration in JSON format, including database and Redis configuration data:

{
  "config.yaml": {
    ....
    "BROWSER_API_CALLS_XHR_ONLY": false,
    "BUILDLOGS_REDIS": {
      "host": "quay-server",
      "password": "strongpassword",
      "port": 6379
    },
    "DATABASE_SECRET_KEY": "4b1c5663-88c6-47ac-b4a8-bb594660f08b",
    "DB_CONNECTION_ARGS": {
      "autorollback": true,
      "threadlocals": true
    },
    "DB_URI": "postgresql://quayuser:quaypass@quay-server:5432/quay",
    "DEFAULT_TAG_EXPIRATION": "2w",
    ....


  }

}

Validating configuration using the API

You can validate a configuration by posting it to the config/validate endpoint:

curl -u quayconfig:secret --header 'Content-Type: application/json' --request POST --data '
{
  "config.yaml": {
    ....
    "BROWSER_API_CALLS_XHR_ONLY": false,
    "BUILDLOGS_REDIS": {
      "host": "quay-server",
      "password": "strongpassword",
      "port": 6379
    },
    "DATABASE_SECRET_KEY": "4b1c5663-88c6-47ac-b4a8-bb594660f08b",
    "DB_CONNECTION_ARGS": {
      "autorollback": true,
      "threadlocals": true
    },
    "DB_URI": "postgresql://quayuser:quaypass@quay-server:5432/quay",
    "DEFAULT_TAG_EXPIRATION": "2w",
    ....

  }

} http://quay-server:8080/api/v1/config/validate | jq

The returned value is an array containing the errors found in the configuration. If the configuration is valid, an empty array [] is returned.

Determining the required fields

You can determine the required fields by posting an empty configuration structure to the config/validate endpoint:

curl -u quayconfig:secret --header 'Content-Type: application/json' --request POST --data '
{
  "config.yaml": {
  }

} http://quay-server:8080/api/v1/config/validate | jq

The value returned is an array indicating which fields are required:

[
  {
    "FieldGroup": "Database",
    "Tags": [
      "DB_URI"
    ],
    "Message": "DB_URI is required."
  },
  {
    "FieldGroup": "DistributedStorage",
    "Tags": [
      "DISTRIBUTED_STORAGE_CONFIG"
    ],
    "Message": "DISTRIBUTED_STORAGE_CONFIG must contain at least one storage location."
  },
  {
    "FieldGroup": "HostSettings",
    "Tags": [
      "SERVER_HOSTNAME"
    ],
    "Message": "SERVER_HOSTNAME is required"
  },
  {
    "FieldGroup": "HostSettings",
    "Tags": [
      "SERVER_HOSTNAME"
    ],
    "Message": "SERVER_HOSTNAME must be of type Hostname"
  },
  {
    "FieldGroup": "Redis",
    "Tags": [
      "BUILDLOGS_REDIS"
    ],
    "Message": "BUILDLOGS_REDIS is required"
  }
]

Getting Project Quay release notifications

To keep up with the latest Project Quay releases and other changes related to Project Quay, you can sign up for update notifications on the Red Hat Customer Portal. After signing up for notifications, you will receive notifications letting you know when there is new a Project Quay version, updated documentation, or other Project Quay news.

  1. Log into the Red Hat Customer Portal with your Red Hat customer account credentials.

  2. Select your user name (upper-right corner) to see Red Hat Account and Customer Portal selections: View account and portal selections

  3. Select Notifications. Your profile activity page appears.

  4. Select the Notifications tab.

  5. Select Manage Notifications.

  6. Select Follow, then choose Products from the drop-down box.

  7. From the drop-down box next to the Products, search for and select Project Quay: Select Products from notifications box

  8. Select the SAVE NOTIFICATION button. Going forward, you will receive notifications when there are changes to the Project Quay product, such as a new release.

Using SSL to protect connections to Project Quay

Introduction to using SSL

To configure Project Quay with a self-signed certificate, you need to create a Certificate Authority (CA) and then generate the required key and certificate files.

The following examples assume you have configured the server hostname quay-server.example.com using DNS or another naming mechanism, such as adding an entry in your /etc/hosts file:

$ cat /etc/hosts
...
192.168.1.112   quay-server.example.com

Create a Certificate Authority and sign a certificate

At the end of this procedure, you will have a certificate file and a primary key file named ssl.cert and ssl.key, respectively.

Create a Certificate Authority

  1. Generate the root CA key:

    $ openssl genrsa -out rootCA.key 2048
  2. Generate the root CA cert:

    $ openssl req -x509 -new -nodes -key rootCA.key -sha256 -days 1024 -out rootCA.pem
  3. Enter the information that will be incorporated into your certificate request, including the server hostname, for example:

    Country Name (2 letter code) [XX]:IE
    State or Province Name (full name) []:GALWAY
    Locality Name (eg, city) [Default City]:GALWAY
    Organization Name (eg, company) [Default Company Ltd]:QUAY
    Organizational Unit Name (eg, section) []:DOCS
    Common Name (eg, your name or your server's hostname) []:quay-server.example.com

Sign a certificate

  1. Generate the server key:

    $ openssl genrsa -out ssl.key 2048
  2. Generate a signing request:

    $ openssl req -new -key ssl.key -out ssl.csr
  3. Enter the information that will be incorporated into your certificate request, including the server hostname, for example:

    Country Name (2 letter code) [XX]:IE
    State or Province Name (full name) []:GALWAY
    Locality Name (eg, city) [Default City]:GALWAY
    Organization Name (eg, company) [Default Company Ltd]:QUAY
    Organizational Unit Name (eg, section) []:DOCS
    Common Name (eg, your name or your server's hostname) []:quay-server.example.com
  4. Create a configuration file openssl.cnf, specifying the server hostname, for example:

    openssl.cnf
    [req]
    req_extensions = v3_req
    distinguished_name = req_distinguished_name
    [req_distinguished_name]
    [ v3_req ]
    basicConstraints = CA:FALSE
    keyUsage = nonRepudiation, digitalSignature, keyEncipherment
    subjectAltName = @alt_names
    [alt_names]
    DNS.1 = quay-server.example.com
    IP.1 = 192.168.1.112
  5. Use the configuration file to generate the certificate ssl.cert:

    $ openssl x509 -req -in ssl.csr -CA rootCA.pem -CAkey rootCA.key -CAcreateserial -out ssl.cert -days 356 -extensions v3_req -extfile openssl.cnf

Configuring SSL using the command line

Another option when configuring SSL is to use the command line interface.

  1. Copy the certificate file and primary key file to your configuration directory, ensuring they are named ssl.cert and ssl.key respectively:

    $ cp ~/ssl.cert $QUAY/config
    $ cp ~/ssl.key $QUAY/config
    $ cd $QUAY/config
  2. Edit the config.yaml file and specify that you want Quay to handle TLS:

    config.yaml
    ...
    SERVER_HOSTNAME: quay-server.example.com
    ...
    PREFERRED_URL_SCHEME: https
    ...
  3. Stop the Quay container and restart the registry:

    $ sudo podman rm -f quay
    $ sudo podman run -d --rm -p 80:8080 -p 443:8443 \
      --name=quay \
      -v $QUAY/config:/conf/stack:Z \
      -v $QUAY/storage:/datastorage:Z \
      quay.io/projectquay/quay:v3.7.1

Configuring SSL using the UI

This section configures SSL using the Quay UI. To configure SSL using the command line interface, see the following section.

  1. Start the Quay container in configuration mode:

    $ sudo podman run --rm -it --name quay_config -p 80:8080 -p 443:8443 quay.io/projectquay/quay:v3.7.1 config secret
  2. In the Server Configuration section, select Project Quay handles TLS for TLS. Upload the certificate file and private key file created earlier, ensuring that the Server Hostname matches the value used when creating the certs. Validate and download the updated configuration.

  3. Stop the Quay container and then restart the registry:

    $ sudo podman rm -f quay
    $ sudo podman run -d --rm -p 80:8080 -p 443:8443 \
    --name=quay \
    -v $QUAY/config:/conf/stack:Z \
    -v $QUAY/storage:/datastorage:Z \
    quay.io/projectquay/quay:v3.7.1

Testing SSL configuration using the command line

  • Use the podman login command to attempt to log in to the Quay registry with SSL enabled:

    $ sudo podman login quay-server.example.com
    Username: quayadmin
    Password:
    
    Error: error authenticating creds for "quay-server.example.com": error pinging docker registry quay-server.example.com: Get "https://quay-server.example.com/v2/": x509: certificate signed by unknown authority
  • Podman does not trust self-signed certificates. As a workaround, use the --tls-verify option:

    $ sudo podman login --tls-verify=false quay-server.example.com
    Username: quayadmin
    Password:
    
    Login Succeeded!

Configuring Podman to trust the root Certificate Authority (CA) is covered in a subsequent section.

Testing SSL configuration using the browser

When you attempt to access the Quay registry, in this case, https://quay-server.example.com, the browser warns of the potential risk:

Potential risk

Proceed to the log in screen, and the browser will notify you that the connection is not secure:

Connection not secure

Configuring the system to trust the root Certificate Authority (CA) is covered in the subsequent section.

Configuring podman to trust the Certificate Authority

Podman uses two paths to locate the CA file, namely, /etc/containers/certs.d/ and /etc/docker/certs.d/.

  • Copy the root CA file to one of these locations, with the exact path determined by the server hostname, and naming the file ca.crt:

    $ sudo cp rootCA.pem /etc/containers/certs.d/quay-server.example.com/ca.crt
  • Alternatively, if you are using Docker, you can copy the root CA file to the equivalent Docker directory:

    $ sudo cp rootCA.pem /etc/docker/certs.d/quay-server.example.com/ca.crt

You should no longer need to use the --tls-verify=false option when logging in to the registry:

$ sudo podman login quay-server.example.com

Username: quayadmin
Password:
Login Succeeded!

Configuring the system to trust the certificate authority

  1. Copy the root CA file to the consolidated system-wide trust store:

    $ sudo cp rootCA.pem /etc/pki/ca-trust/source/anchors/
  2. Update the system-wide trust store configuration:

    $ sudo update-ca-trust extract
  3. You can use the trust list command to ensure that the Quay server has been configured:

    $ trust list | grep quay
        label: quay-server.example.com

    Now, when you browse to the registry at https://quay-server.example.com, the lock icon shows that the connection is secure:

    Connection not secure

  4. To remove the root CA from system-wide trust, delete the file and update the configuration:

    $ sudo rm /etc/pki/ca-trust/source/anchors/rootCA.pem
    $ sudo update-ca-trust extract
    $ trust list | grep quay
    $

More information can be found in the RHEL 8 documentation in the chapter Using shared system certificates.

Adding TLS Certificates to the Project Quay Container

To add custom TLS certificates to Project Quay, create a new directory named extra_ca_certs/ beneath the Project Quay config directory. Copy any required site-specific TLS certificates to this new directory.

Add TLS certificates to Project Quay

  1. View certificate to be added to the container

    $ cat storage.crt
    -----BEGIN CERTIFICATE-----
    MIIDTTCCAjWgAwIBAgIJAMVr9ngjJhzbMA0GCSqGSIb3DQEBCwUAMD0xCzAJBgNV
    [...]
    -----END CERTIFICATE-----
  2. Create certs directory and copy certificate there

    $ mkdir -p quay/config/extra_ca_certs
    $ cp storage.crt quay/config/extra_ca_certs/
    $ tree quay/config/
    ├── config.yaml
    ├── extra_ca_certs
    │   ├── storage.crt
  3. Obtain the Quay container’s CONTAINER ID with podman ps:

    $ sudo podman ps
    CONTAINER ID        IMAGE                                COMMAND                  CREATED             STATUS              PORTS
    5a3e82c4a75f        <registry>/<repo>/quay:v3.7.1 "/sbin/my_init"          24 hours ago        Up 18 hours         0.0.0.0:80->80/tcp, 0.0.0.0:443->443/tcp, 443/tcp   grave_keller
  4. Restart the container with that ID:

    $ sudo podman restart 5a3e82c4a75f
  5. Examine the certificate copied into the container namespace:

    $ sudo podman exec -it 5a3e82c4a75f cat /etc/ssl/certs/storage.pem
    -----BEGIN CERTIFICATE-----
    MIIDTTCCAjWgAwIBAgIJAMVr9ngjJhzbMA0GCSqGSIb3DQEBCwUAMD0xCzAJBgNV

Add certs when deployed on Kubernetes

When deployed on Kubernetes, Project Quay mounts in a secret as a volume to store config assets. Unfortunately, this currently breaks the upload certificate function of the superuser panel.

To get around this error, a base64 encoded certificate can be added to the secret after Project Quay has been deployed. Here’s how:

  1. Begin by base64 encoding the contents of the certificate:

    $ cat ca.crt
    -----BEGIN CERTIFICATE-----
    MIIDljCCAn6gAwIBAgIBATANBgkqhkiG9w0BAQsFADA5MRcwFQYDVQQKDA5MQUIu
    TElCQ09SRS5TTzEeMBwGA1UEAwwVQ2VydGlmaWNhdGUgQXV0aG9yaXR5MB4XDTE2
    MDExMjA2NTkxMFoXDTM2MDExMjA2NTkxMFowOTEXMBUGA1UECgwOTEFCLkxJQkNP
    UkUuU08xHjAcBgNVBAMMFUNlcnRpZmljYXRlIEF1dGhvcml0eTCCASIwDQYJKoZI
    [...]
    -----END CERTIFICATE-----
    
    $ cat ca.crt | base64 -w 0
    [...]
    c1psWGpqeGlPQmNEWkJPMjJ5d0pDemVnR2QNCnRsbW9JdEF4YnFSdVd3PT0KLS0tLS1FTkQgQ0VSVElGSUNBVEUtLS0tLQo=
  2. Use the kubectl tool to edit the quay-enterprise-config-secret.

    $ kubectl --namespace quay-enterprise edit secret/quay-enterprise-config-secret
  3. Add an entry for the cert and paste the full base64 encoded string under the entry:

      custom-cert.crt:
    c1psWGpqeGlPQmNEWkJPMjJ5d0pDemVnR2QNCnRsbW9JdEF4YnFSdVd3PT0KLS0tLS1FTkQgQ0VSVElGSUNBVEUtLS0tLQo=
  4. Finally, recycle all Project Quay pods. Use kubectl delete to remove all Project Quay pods. The Project Quay Deployment will automatically schedule replacement pods with the new certificate data.

Configuring action log storage for Elasticsearch

By default, the past three months of usage logs are stored in the Project Quay database and exposed via the web UI on organization and repository levels. Appropriate administrative privileges are required to see log entries. For deployments with a large amount of logged operations, you can now store the usage logs in Elasticsearch instead of the Project Quay database backend. To do this, you need to provide your own Elasticsearch stack, as it is not included with Project Quay as a customizable component.

Enabling Elasticsearch logging can be done during Project Quay deployment or post-deployment using the Project Quay Config Tool. The resulting configuration is stored in the config.yaml file. Once configured, usage log access continues to be provided the same way, via the web UI for repositories and organizations.

Here’s how to configure action log storage to change it from the default Project Quay database to use Elasticsearch:

  1. Obtain an Elasticsearch account.

  2. Open the Project Quay Config Tool (either during or after Project Quay deployment).

  3. Scroll to the Action Log Storage Configuration setting and select Elasticsearch instead of Database. The following figure shows the Elasticsearch settings that appear:

    Choose Elasticsearch to view settings to store logs

  4. Fill in the following information for your Elasticsearch instance:

    • Elasticsearch hostname: The hostname or IP address of the system providing the Elasticsearch service.

    • Elasticsearch port: The port number providing the Elasticsearch service on the host you just entered. Note that the port must be accessible from all systems running the Project Quay registry. The default is TCP port 9200.

    • Elasticsearch access key: The access key needed to gain access to the Elastic search service, if required.

    • Elasticsearch secret key: The secret key needed to gain access to the Elastic search service, if required.

    • AWS region: If you are running on AWS, set the AWS region (otherwise, leave it blank).

    • Index prefix: Choose a prefix to attach to log entries.

    • Logs Producer: Choose either Elasticsearch (default) or Kinesis to direct logs to an intermediate Kinesis stream on AWS. You need to set up your own pipeline to send logs from Kinesis to Elasticsearch (for example, Logstash). The following figure shows additional fields you would need to fill in for Kinesis:

      On AWS optionally set up an intermediate Kinesis stream

  5. If you chose Elasticsearch as the Logs Producer, no further configuration is needed. If you chose Kinesis, fill in the following:

    • Stream name: The name of the Kinesis stream.

    • AWS access key: The name of the AWS access key needed to gain access to the Kinesis stream, if required.

    • AWS secret key: The name of the AWS secret key needed to gain access to the Kinesis stream, if required.

    • AWS region: The AWS region.

  6. When you are done, save the configuration. The Config Tool checks your settings. If there is a problem connecting to the Elasticsearch or Kinesis services, you will see an error and have the opportunity to continue editing. Otherwise, logging will begin to be directed to your Elasticsearch configuration after the cluster restarts with the new configuration.

Clair Security Scanning

Clair is a set of micro services that can be used with Project Quay to perform vulnerability scanning of container images associated with a set of Linux operating systems. The micro services design of Clair makes it appropriate to run in a highly scalable configuration, where components can be scaled separately as appropriate for enterprise environments.

Clair uses the following vulnerability databases to scan for issues in your images:

  • Alpine SecDB database

  • AWS UpdateInfo

  • Debian Oval database

  • Oracle Oval database

  • RHEL Oval database

  • SUSE Oval database

  • Ubuntu Oval database

  • Pyup.io (python) database

For information on how Clair does security mapping with the different databases, see ClairCore Severity Mapping.

Note

With the release of Clair V4 (image clair), the previously used Clair V2 (image clair-jwt) is no longer used. See below for how to run V2 in read-only mode while V4 is updating.

Setting Up Clair on a Project Quay OpenShift deployment

Deploying Via the Quay Operator

To set up Clair V4 on a new Project Quay deployment on OpenShift, it is highly recommended to use the Quay Operator. By default, the Quay Operator will install or upgrade a Clair deployment along with your Project Quay deployment and configure Clair security scanning automatically.

Manually Deploying Clair

To configure Clair V4 on an existing Project Quay OpenShift deployment running Clair V2, first ensure Project Quay has been upgraded to at least version 3.4.0. Then use the following steps to manually set up Clair V4 alongside Clair V2.

  1. Set your current project to the name of the project in which Project Quay is running. For example:

    $ oc project quay-enterprise
  2. Create a Postgres deployment file for Clair v4 (for example, clairv4-postgres.yaml) as follows.

    clairv4-postgres.yaml
    ---
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: clairv4-postgres
      namespace: quay-enterprise
      labels:
        quay-component: clairv4-postgres
    spec:
      replicas: 1
      selector:
        matchLabels:
          quay-component: clairv4-postgres
      template:
        metadata:
          labels:
            quay-component: clairv4-postgres
        spec:
          volumes:
            - name: postgres-data
              persistentVolumeClaim:
                claimName: clairv4-postgres
          containers:
            - name: postgres
              image: postgres:11.5
              imagePullPolicy: "IfNotPresent"
              ports:
                - containerPort: 5432
              env:
                - name: POSTGRES_USER
                  value: "postgres"
                - name: POSTGRES_DB
                  value: "clair"
                - name: POSTGRES_PASSWORD
                  value: "postgres"
                - name: PGDATA
                  value: "/etc/postgres/data"
              volumeMounts:
                - name: postgres-data
                  mountPath: "/etc/postgres"
    ---
    apiVersion: v1
    kind: PersistentVolumeClaim
    metadata:
      name: clairv4-postgres
      labels:
        quay-component: clairv4-postgres
    spec:
      accessModes:
        - "ReadWriteOnce"
      resources:
        requests:
          storage: "5Gi"
        volumeName: "clairv4-postgres"
    ---
    apiVersion: v1
    kind: Service
    metadata:
      name: clairv4-postgres
      labels:
        quay-component: clairv4-postgres
    spec:
      type: ClusterIP
      ports:
        - port: 5432
          protocol: TCP
          name: postgres
          targetPort: 5432
      selector:
        quay-component: clairv4-postgres
  3. Deploy the postgres database as follows:

    $ oc create -f ./clairv4-postgres.yaml
  4. Create a Clair config.yaml file to use for Clair v4. For example:

    config.yaml
    introspection_addr: :8089
    http_listen_addr: :8080
    log_level: debug
    indexer:
      connstring: host=clairv4-postgres port=5432 dbname=clair user=postgres password=postgres sslmode=disable
      scanlock_retry: 10
      layer_scan_concurrency: 5
      migrations: true
    matcher:
      connstring: host=clairv4-postgres port=5432 dbname=clair user=postgres password=postgres sslmode=disable
      max_conn_pool: 100
      run: ""
      migrations: true
      indexer_addr: clair-indexer
    notifier:
      connstring: host=clairv4-postgres port=5432 dbname=clair user=postgres password=postgres sslmode=disable
      delivery: 1m
      poll_interval: 5m
      migrations: true
    auth:
      psk:
        key: MTU5YzA4Y2ZkNzJoMQ== (1)
        iss: ["quay"]
    # tracing and metrics
    trace:
      name: "jaeger"
      probability: 1
      jaeger:
        agent_endpoint: "localhost:6831"
        service_name: "clair"
    metrics:
      name: "prometheus"
    1. To generate a Clair pre-shared key (PSK), enable scanning in the Security Scanner section of the User Interface and click Generate PSK.

More information about Clair’s configuration format can be found in upstream Clair documentation.

  1. Create a secret from the Clair config.yaml:

    $ oc create secret generic clairv4-config-secret --from-file=./config.yaml
  2. Create the Clair v4 deployment file (for example, clair-combo.yaml) and modify it as necessary:

    clair-combo.yaml
    ---
    apiVersion: extensions/v1beta1
    kind: Deployment
    metadata:
      labels:
        quay-component: clair-combo
      name: clair-combo
    spec:
      replicas: 1
      selector:
        matchLabels:
          quay-component: clair-combo
      template:
        metadata:
          labels:
            quay-component: clair-combo
        spec:
          containers:
            - image: quay.io/projectquay/clair:v3.7.1  (1)
              imagePullPolicy: IfNotPresent
              name: clair-combo
              env:
                - name: CLAIR_CONF
                  value: /clair/config.yaml
                - name: CLAIR_MODE
                  value: combo
              ports:
                - containerPort: 8080
                  name: clair-http
                  protocol: TCP
                - containerPort: 8089
                  name: clair-intro
                  protocol: TCP
              volumeMounts:
                - mountPath: /clair/
                  name: config
          imagePullSecrets:
            - name: redhat-pull-secret
          restartPolicy: Always
          volumes:
            - name: config
              secret:
                secretName: clairv4-config-secret
    ---
    apiVersion: v1
    kind: Service
    metadata:
      name: clairv4 (2)
      labels:
        quay-component: clair-combo
    spec:
      ports:
        - name: clair-http
          port: 80
          protocol: TCP
          targetPort: 8080
        - name: clair-introspection
          port: 8089
          protocol: TCP
          targetPort: 8089
      selector:
        quay-component: clair-combo
      type: ClusterIP
    1. Change image to latest clair image name and version.

    2. With the Service set to clairv4, the scanner endpoint for Clair v4 is entered later into the Project Quay config.yaml in the SECURITY_SCANNER_V4_ENDPOINT as http://clairv4.

  3. Create the Clair v4 deployment as follows:

    $ oc create -f ./clair-combo.yaml
  4. Modify the config.yaml file for your Project Quay deployment to add the following entries at the end:

    FEATURE_SECURITY_NOTIFICATIONS: true
    FEATURE_SECURITY_SCANNER: true
    SECURITY_SCANNER_V4_ENDPOINT: http://clairv4 (1)
    1. Identify the Clair v4 service endpoint

  5. Redeploy the modified config.yaml to the secret containing that file (for example, quay-enterprise-config-secret:

    $ oc delete secret quay-enterprise-config-secret
    $ oc create secret generic quay-enterprise-config-secret --from-file=./config.yaml
  6. For the new config.yaml to take effect, you need to restart the Project Quay pods. Simply deleting the quay-app pods causes pods with the updated configuration to be deployed.

At this point, images in any of the organizations identified in the namespace whitelist will be scanned by Clair v4.

Setting up Clair on a non-OpenShift Project Quay deployment

For Project Quay deployments not running on OpenShift, it is possible to configure Clair security scanning manually. Project Quay deployments already running Clair V2 can use the instructions below to add Clair V4 to their deployment.

  1. Deploy a (preferably fault-tolerant) Postgres database server. Note that Clair requires the uuid-ossp extension to be added to its Postgres database. If the user supplied in Clair’s config.yaml has the necessary privileges to create the extension then it will be added automatically by Clair itself. If not, then the extension must be added before starting Clair. If the extension is not present, the following error will be displayed when Clair attempts to start.

    ERROR: Please load the "uuid-ossp" extension. (SQLSTATE 42501)
  2. Create a Clair config file in a specific folder, for example, /etc/clairv4/config/config.yaml).

    config.yaml
    introspection_addr: :8089
    http_listen_addr: :8080
    log_level: debug
    indexer:
      connstring: host=clairv4-postgres port=5432 dbname=clair user=postgres password=postgres sslmode=disable
      scanlock_retry: 10
      layer_scan_concurrency: 5
      migrations: true
    matcher:
      connstring: host=clairv4-postgres port=5432 dbname=clair user=postgres password=postgres sslmode=disable
      max_conn_pool: 100
      run: ""
      migrations: true
      indexer_addr: clair-indexer
    notifier:
      connstring: host=clairv4-postgres port=5432 dbname=clair user=postgres password=postgres sslmode=disable
      delivery_interval: 1m
      poll_interval: 5m
      migrations: true
    
    # tracing and metrics
    trace:
      name: "jaeger"
      probability: 1
      jaeger:
        agent_endpoint: "localhost:6831"
        service_name: "clair"
    metrics:
      name: "prometheus"

More information about Clair’s configuration format can be found in upstream Clair documentation.

  1. Run Clair via the container image, mounting in the configuration from the file you created.

    $ podman run -p 8080:8080 -p 8089:8089 -e CLAIR_CONF=/clair/config.yaml -e CLAIR_MODE=combo -v /etc/clair4/config:/clair -d quay.io/projectquay/clair:v3.7.1
  2. Follow the remaining instructions from the previous section for configuring Project Quay to use the new Clair V4 endpoint.

Running multiple Clair containers in this fashion is also possible, but for deployment scenarios beyond a single container the use of a container orchestrator like Kubernetes or OpenShift is strongly recommended.

Advanced Clair configuration

Unmanaged Clair configuration

With Project Quay 3.7, users can run an unmanaged Clair configuration on the Project Quay OpenShift Container Platform Operator. This feature allows users to create an unmanaged Clair database, or run their custom Clair configuration without an unmanaged database.

Unmanaging a Clair database

An unmanaged Clair database allows the Project Quay Operator to work in a geo-replicated environment, where multiple instances of the Operator must communicate with the same database. An unmanaged Clair database can also be used when a user requires a highly-available (HA) Clair database that exists outside of a cluster.

Procedure
  • In the Quay Operator, set the clairpostgres component of the QuayRegistry custom resource to unmanaged:

    apiVersion: quay.redhat.com/v1
    kind: QuayRegistry
    metadata:
      name: quay370
    spec:
      configBundleSecret: config-bundle-secret
      components:
        - kind: objectstorage
          managed: false
        - kind: route
          managed: true
        - kind: tls
          managed: false
        - kind: clairpostgres
          managed: false
Configuring a custom Clair database

The Project Quay Operator for OpenShift Container Platform allows users to provide their own Clair configuration by editing the configBundleSecret parameter.

Procedure
  1. Create a Quay config bundle secret that includes the clair-config.yaml:

    $ oc create secret generic --from-file config.yaml=./config.yaml --from-file extra_ca_cert_rds-ca-2019-root.pem=./rds-ca-2019-root.pem --from-file clair-config.yaml=./clair-config.yaml --from-file ssl.cert=./ssl.cert --from-file ssl.key=./ssl.key config-bundle-secret

    Example clair-config.yaml configuration:

    indexer:
        connstring: host=quay-server.example.com port=5432 dbname=quay user=quayrdsdb password=quayrdsdb sslrootcert=/run/certs/rds-ca-2019-root.pem sslmode=verify-ca
        layer_scan_concurrency: 6
        migrations: true
        scanlock_retry: 11
    log_level: debug
    matcher:
        connstring: host=quay-server.example.com port=5432 dbname=quay user=quayrdsdb password=quayrdsdb sslrootcert=/run/certs/rds-ca-2019-root.pem sslmode=verify-ca
        migrations: true
    metrics:
        name: prometheus
    notifier:
        connstring: host=quay-server.example.com port=5432 dbname=quay user=quayrdsdb password=quayrdsdb sslrootcert=/run/certs/rds-ca-2019-root.pem sslmode=verify-ca
        migrations: true
    Note
    • The database certificate is mounted under /run/certs/rds-ca-2019-root.pem on the Clair application pod in the clair-config.yaml. It must be specified when configuring your clair-config.yaml.

    • An example clair-config.yaml can be found at Clair on OpenShift config.

  2. Add the clair-config.yaml to your bundle secret, named configBundleSecret:

    apiVersion: v1
    kind: Secret
    metadata:
      name: config-bundle-secret
      namespace: quay-enterprise
    data:
      config.yaml: <base64 encoded Quay config>
      clair-config.yaml: <base64 encoded Clair config>
      extra_ca_cert_<name>: <base64 encoded ca cert>
      clair-ssl.crt: >-
      clair-ssl.key: >-
    Note

    When updated, the provided clair-config.yaml is mounted into the Clair pod. Any fields not provided are automatically populated with defaults using the Clair configuration module.

After proper configuration, the Clair application pod should return to a Ready state.

Running a custom Clair configuration with a managed database

In some cases, users might want to run a custom Clair configuration with a managed database. This is useful in the following scenarios:

  • When a user wants to disable an updater.

  • When a user is running in an air-gapped environment.

    Note
    • If you are running Quay in an air-gapped environment, the airgap parameter of your clair-config.yaml must be set to true.

    • If you are running Quay in an air-gapped environment, you should disable all updaters.

Use the steps in "Configuring a custom Clair database" to configure your database when clairpostgres is set to managed.

For more information about running Clair in an air-gapped environment, see Configuring access to the Clair database in the air-gapped OpenShift cluster.

Clair CRDA configuration

Enabling Clair CRDA

With Project Quay 3.7, Java scanning no longer includes a default CRDA shared key, and is no longer enabled by default. Use the following procedure to fetch the Quay-specific CRDA remote matcher, which supports higher RPS, and to manually enable CRDA for Java scanning.

Prerequisites
  • Project Quay 3.7

Procedure
  1. Submit the API key request form to obtain the Quay-specific CRDA remote matcher.

  2. Set the CRDA configuration in your clair-config.yaml file:

    ```
    matchers:
           crda:
             url: https://f8a-analytics-2445582058137.production.gw.apicast.io/api/v2/
             source: quay.io
             key: 9e7da76708fe374d8c10fa752e72989f
    ```

Using Clair

  1. Log in to your Project Quay cluster and select an organization for which you have configured Clair scanning.

  2. Select a repository from that organization that holds some images and select Tags from the left navigation. The following figure shows an example of a repository with two images that have been scanned:

    Security scan information appears for scanned repository images

  3. If vulnerabilities are found, select to under the Security Scan column for the image to see either all vulnerabilities or those that are fixable. The following figure shows information on all vulnerabilities found:

    See all vulnerabilities or only those that are fixable

CVE ratings from the National Vulnerability Database

With Clair v4.2, enrichment data is now viewable in the Quay UI. Additionally, Clair v4.2 adds CVSS scores from the National Vulnerability Database for detected vulnerabilities.

With this change, if the vulnerability has a CVSS score that is within 2 levels of the distro’s score, the Quay UI present’s the distro’s score by default. For example:

Clair v4.2 data display

This differs from the previous interface, which would only display the following information:

Clair v4 data display

Configuring Clair for Disconnected Environments

Clair utilizes a set of components called Updaters to handle the fetching and parsing of data from various vulnerability databases. These Updaters are set up by default to pull vulnerability data directly from the internet and work out of the box. For customers in disconnected environments without direct access to the internet this poses a problem. Clair supports these environments through the ability to work with different types of update workflows that take into account network isolation. Using the clairctl command line utility, any process can easily fetch Updater data from the internet via an open host, securely transfer the data to an isolated host, and then import the Updater data on the isolated host into Clair itself.

The steps are as follows.

  1. First ensure that your Clair configuration has disabled automated Updaters from running.

    config.yaml
    matcher:
      disable_updaters: true
  2. Export out the latest Updater data to a local archive. This requires the clairctl tool which can be run directly as a binary, or via the Clair container image. Assuming your Clair configuration is in /etc/clairv4/config/config.yaml, to run via the container image:

    $ podman run -it --rm -v /etc/clairv4/config:/cfg:Z -v /path/to/output/directory:/updaters:Z --entrypoint /bin/clairctl quay.io/projectquay/clair:v3.7.1 --config /cfg/config.yaml export-updaters  /updaters/updaters.gz

    Note that you need to explicitly reference the Clair configuration. This will create the Updater archive in /etc/clairv4/updaters/updaters.gz. If you want to ensure the archive was created without any errors from the source databases, you can supply the --strict flag to clairctl. The archive file should be copied over to a volume that is accessible from the disconnected host running Clair. From the disconnected host, use the same procedure now to import the archive into Clair.

    $ podman run -it --rm -v /etc/clairv4/config:/cfg:Z -v /path/to/output/directory:/updaters:Z --entrypoint /bin/clairctl quay.io/projectquay/clair:v3.7.1 --config /cfg/config.yaml import-updaters /updaters/updaters.gz

Mapping repositories to Common Product Enumeration (CPE) information

Clair’s RHEL scanner relies on a Common Product Enumeration (CPE) file to properly map RPM packages to the corresponding security data, in order to produce matching results. This file must be present, or access to the file must be allowed, for the scanner to properly process RPMs. If the file is not present, RPMs installed in the container images will not be scanned.

Red Hat publishes the JSON mapping file at https://www.redhat.com/security/data/metrics/repository-to-cpe.json.

In addition to uploading CVE information to the database for disconnected Clair, you must also make the mapping file available locally:

  • For standalone Quay and Clair deployments, the mapping file must be loaded into the Clair pod.

  • For Operator-based deployments, you must set the Clair component to unmanaged. Then deploy Clair manually, setting the configuration to load a local copy of the mapping file.

Use the repo2cpe_mapping_file field in the Clair configuration to specify the file:

indexer:
  scanner:
    repo:
      rhel-repository-scanner:
        repo2cpe_mapping_file: /path/to/repository-to-cpe.json

Further information is available from Red Hat at How to accurately match OVAL security data to installed RPMs.

Clair updater URLs

The following are the HTTP hosts and paths that Clair will attempt to talk to in a default configuration. This list is non-exhaustive, as some servers will issue redirects and some request URLs are constructed dynamically.

  • https://secdb.alpinelinux.org/

  • http://repo.us-west-2.amazonaws.com/2018.03/updates/x86_64/mirror.list

  • https://cdn.amazonlinux.com/2/core/latest/x86_64/mirror.list

  • https://www.debian.org/security/oval/

  • https://linux.oracle.com/security/oval/

  • https://packages.vmware.com/photon/photon_oval_definitions/

  • https://github.com/pyupio/safety-db/archive/

  • https://catalog.redhat.com/api/containers/

  • https://www.redhat.com/security/data/

  • https://support.novell.com/security/oval/

  • https://people.canonical.com/~ubuntu-security/oval/

Additional Information

For detailed documentation on the internals of Clair, including how the microservices are structured, please see the Upstream Clair and ClairCore documentation.

Scan pod images with the Container Security Operator

Using the Container Security Operator, (CSO) you can scan container images associated with active pods, running on OpenShift (4.2 or later) and other Kubernetes platforms, for known vulnerabilities. The CSO:

  • Watches containers associated with pods on all or specified namespaces

  • Queries the container registry where the containers came from for vulnerability information provided an image’s registry supports image scanning (such as a Quay registry with Clair scanning)

  • Exposes vulnerabilities via the ImageManifestVuln object in the Kubernetes API

Using the instructions here, the CSO is installed in the marketplace-operators namespace, so it is available to all namespaces on your OpenShift cluster.

Note

To see instructions on installing the CSO on Kubernetes, select the Install button from the Container Security OperatorHub.io page.

Run the CSO in OpenShift

To start using the CSO in OpenShift, do the following:

  1. Go to Operators → OperatorHub (select Security) to see the available Container Security Operator.

  2. Select the Container Security Operator, then select Install to go to the Create Operator Subscription page.

  3. Check the settings (all namespaces and automatic approval strategy, by default), and select Subscribe. The Container Security appears after a few moments on the Installed Operators screen.

  4. Optionally, you can add custom certificates to the CSO. In this example, create a certificate named quay.crt in the current directory. Then run the following command to add the cert to the CSO (restart the Operator pod for the new certs to take effect):

    $ oc create secret generic container-security-operator-extra-certs --from-file=quay.crt -n openshift-operators
  5. Open the OpenShift Dashboard (Home → Dashboards). A link to Image Security appears under the status section, with a listing of the number of vulnerabilities found so far. Select the link to see a Security breakdown, as shown in the following figure:

    Access SCO scanning data from OpenShift dashboard

  6. You can do one of two things at this point to follow up on any detected vulnerabilities:

    • Select the link to the vulnerability. You are taken to the container registry, Project Quay or other registry where the container came from, where you can see information about the vulnerability. The following figure shows an example of detected vulnerabilities from a Quay.io registry:

      The CSO points you to a registry containing the vulnerable image

    • Select the namespaces link to go to the ImageManifestVuln screen, where you can see the name of the selected image and all namespaces where that image is running. The following figure indicates that a particular vulnerable image is running in two namespaces:

      View namespaces a vulnerable image is running in

At this point, you know what images are vulnerable, what you need to do to fix those vulnerabilities, and every namespace that the image was run in. So you can:

  • Alert anyone running the image that they need to correct the vulnerability

  • Stop the images from running (by deleting the deployment or other object that started the pod the image is in)

Note that if you do delete the pod, it may take a few minutes for the vulnerability to reset on the dashboard.

Query image vulnerabilities from the CLI

You can query information on security from the command line. To query for detected vulnerabilities, type:

$ oc get vuln --all-namespaces
NAMESPACE     NAME              AGE
default       sha256.ca90...    6m56s
skynet        sha256.ca90...    9m37s

To display details for a particular vulnerability, identify one of the vulnerabilities, along with its namespace and the describe option. This example shows an active container whose image includes an RPM package with a vulnerability:

$ oc describe vuln --namespace mynamespace sha256.ac50e3752...
Name:         sha256.ac50e3752...
Namespace:    quay-enterprise
...
Spec:
  Features:
    Name:            nss-util
    Namespace Name:  centos:7
    Version:         3.44.0-3.el7
    Versionformat:   rpm
    Vulnerabilities:
      Description: Network Security Services (NSS) is a set of libraries...

Project Quay Security Scanning with Clair V2

Project Quay supports scanning container images for known vulnerabilities with a scanning engine such as Clair. This document explains how to configure Clair with Project Quay.

Note

With the release of Project Quay 3.4, the default version of Clair is V4. This new version V4 is no longer being released as Technology Preview and is supported for production use. Customers are strongly encouraged to use Clair V4 for with Project Quay 3.4. It is possible to run both Clair V4 and Clair V2 simultaneously if so desired. In future versions of Project Quay, Clair V2 will eventually be removed.

Set up Clair V2 in the Project Quay config tool

Enabling Clair V2 in Project Quay consists of:

  • Starting the Project Quay config tool. See the Project Quay deployment guide for the type of deployment you are doing (OpenShift, Basic, or HA) for how to start the config tool for that environment.

  • Enabling security scanning, then generating a private key and PEM file in the config tool

  • Including the key and PEM file in the Clair config file

  • Start the Clair container

The procedure varies, based on whether you are running Project Quay on OpenShift or directly on a host.

Enabling Clair V2 on a Project Quay OpenShift deployment

To set up Clair V2 on Project Quay in OpenShift, see Add Clair image scanning to Project Quay.

Enabling Clair V2 on a Project Quay Basic or HA deployment

To set up Clair V2 on a Project Quay deployment where the container is running directly on the host system, do the following:

  1. Restart the Project Quay config tool: Run the Quay container again in config mode, open the configuration UI in a browser, then select Modify an existing configuration. When prompted, upload the quay-config.tar.gz file that was originally created for the deployment.

  2. Enable Security Scanning: Scroll to the Security Scanner section and select the "Enable Security Scanning" checkbox. From the fields that appear you need to create an authentication key and enter the security scanner endpoint. Here’s how:

    • Generate key: Click Create Key, then from the pop-up window type a name for the Clair private key and an optional expiration date (if blank, the key never expires). Then select Generate Key.

    • Copy the Clair key and PEM file: Save the Key ID (to a notepad or similar) and download a copy of the Private Key PEM file (named security_scanner.pem) by selecting "Download Private Key" (if you lose the key, you need to generate a new one). You will need the key and PEM file when you start the Clair container later.

      Close the pop-up when you are done. Here is an example of a completed Security Scanner config:

      Create authentication key and set scan endpoint

  3. Save the configuration: Click Save Configuration Changes and then select Download Configuration to save it to your local system.

  4. Deploy the configuration: To pick up the changes enabling scanning, as well as other changes you may have made to the configuration, unpack the quay-config.tar.gz and copy the resulting files to the config directory. For example:

    $ tar xvf quay-config.tar.gz
    config.yaml  ssl.cert  ssl.key
    $ cp config.yaml ssl* /mnt/quay/config

Next, start the Clair V2 container and associated database, as described in the following sections.

Setting Up Clair V2 Security Scanning

Once you have created the necessary key and pem files from the Project Quay config UI, you are ready to start up the Clair V2 container and associated database. Once that is done, you an restart your Project Quay cluster to have those changes take effect.

Procedures for running the Clair V2 container and associated database are different on OpenShift than they are for running those containers directly on a host.

Run Clair V2 on a Project Quay OpenShift deployment

To run the Clair V2 image scanning container and its associated database on an OpenShift environment with your Project Quay cluster, see Add Clair image scanning to Project Quay.

Run Clair V2 on a Project Quay Basic or HA deployment

To run Clair V2 and its associated database on non-OpenShift environments (directly on a host), you need to:

  • Start up a database

  • Configure and start Clair V2

Get Postgres and Clair

In order to run Clair, a database is required. For production deployments, MySQL is not supported. For production, we recommend you use PostgreSQL or other supported database:

  • Running on machines other than those running Project Quay

  • Ideally with automatic replication and failover

For testing purposes, a single PostgreSQL instance can be started locally:

  1. To start Postgres locally, do the following:

    # sudo podman run --name postgres -p 5432:5432 -d postgres
    # sleep 5
    # sudo podman run --rm --link postgres:postgres postgres \
       sh -c 'echo "create database clairtest" | psql -h \
       "$POSTGRES_PORT_5432_TCP_ADDR" -p  \
       "$POSTGRES_PORT_5432_TCP_PORT" -U postgres'

    The configuration string for this test database is:

    postgresql://postgres@{DOCKER HOST GOES HERE}:5432/clairtest?sslmode=disable
  2. Pull the security-enabled Clair image:

You will need to build your own Clair container and pull it during this step. Instructions for building the Clair container are not yet available.

  1. Make a configuration directory for Clair

    # mkdir clair-config
    # cd clair-config

Configure Clair V2

Clair V2 can run either as a single instance or in high-availability mode. It is recommended to run more than a single instance of Clair, ideally in an auto-scaling group with automatic healing.

  1. Create a config.yaml file to be used in the Clair V2 config directory (/clair/config) from one of the two Clair configuration files shown here.

  2. If you are doing a high-availability installation, go through the procedure in Authentication for high-availability scanners to create a Key ID and Private Key (PEM).

  3. Save the Private Key (PEM) to a file (such as, $HOME/config/security_scanner.pem).

  4. Replace the value of key_id (CLAIR_SERVICE_KEY_ID) with the Key ID you generated and the value of private_key_path with the location of the PEM file (for example, /config/security_scanner.pem).

    For example, those two value might now appear as:

    key_id: { 4fb9063a7cac00b567ee921065ed16fed7227afd806b4d67cc82de67d8c781b1 }
    private_key_path: /clair/config/security_scanner.pem
  5. Change other values in the configuration file as needed.

Clair V2 configuration: High availability
clair:
  database:
    type: pgsql
    options:
      # A PostgreSQL Connection string pointing to the Clair Postgres database.
      # Documentation on the format can be found at: http://www.postgresql.org/docs/9.4/static/libpq-connect.html
      source: { POSTGRES_CONNECTION_STRING }
      cachesize: 16384
  api:
    # The port at which Clair will report its health status. For example, if Clair is running at
    # https://clair.mycompany.com, the health will be reported at
    # http://clair.mycompany.com:6061/health.
    healthport: 6061

    port: 6062
    timeout: 900s

    # paginationkey can be any random set of characters. *Must be the same across all Clair instances*.
    paginationkey: "XxoPtCUzrUv4JV5dS+yQ+MdW7yLEJnRMwigVY/bpgtQ="

  updater:
    # interval defines how often Clair will check for updates from its upstream vulnerability databases.
    interval: 6h
  notifier:
    attempts: 3
    renotifyinterval: 1h
    http:
      # QUAY_ENDPOINT defines the endpoint at which Quay is running.
      # For example: https://myregistry.mycompany.com
      endpoint: { QUAY_ENDPOINT }/secscan/notify
      proxy: http://localhost:6063

jwtproxy:
  signer_proxy:
    enabled: true
    listen_addr: :6063
    ca_key_file: /certificates/mitm.key # Generated internally, do not change.
    ca_crt_file: /certificates/mitm.crt # Generated internally, do not change.
    signer:
      issuer: security_scanner
      expiration_time: 5m
      max_skew: 1m
      nonce_length: 32
      private_key:
        type: preshared
        options:
          # The ID of the service key generated for Clair. The ID is returned when setting up
          # the key in [Quay Setup](security-scanning.md)
          key_id: { CLAIR_SERVICE_KEY_ID }
          private_key_path: /clair/config/security_scanner.pem

  verifier_proxies:
  - enabled: true
    # The port at which Clair will listen.
    listen_addr: :6060

    # If Clair is to be served via TLS, uncomment these lines. See the "Running Clair under TLS"
    # section below for more information.
    # key_file: /clair/config/clair.key
    # crt_file: /clair/config/clair.crt

    verifier:
      # CLAIR_ENDPOINT is the endpoint at which this Clair will be accessible. Note that the port
      # specified here must match the listen_addr port a few lines above this.
      # Example: https://myclair.mycompany.com:6060
      audience: { CLAIR_ENDPOINT }

      upstream: http://localhost:6062
      key_server:
        type: keyregistry
        options:
          # QUAY_ENDPOINT defines the endpoint at which Quay is running.
          # Example: https://myregistry.mycompany.com
          registry: { QUAY_ENDPOINT }/keys/
Clair V2 configuration: Single instance
clair:
  database:
    type: pgsql
    options:
      # A PostgreSQL Connection string pointing to the Clair Postgres database.
      # Documentation on the format can be found at: http://www.postgresql.org/docs/9.4/static/libpq-connect.html
      source: { POSTGRES_CONNECTION_STRING }
      cachesize: 16384
  api:
    # The port at which Clair will report its health status. For example, if Clair is running at
    # https://clair.mycompany.com, the health will be reported at
    # http://clair.mycompany.com:6061/health.
    healthport: 6061

    port: 6062
    timeout: 900s

    # paginationkey can be any random set of characters. *Must be the same across all Clair instances*.
    paginationkey:

  updater:
    # interval defines how often Clair will check for updates from its upstream vulnerability databases.
    interval: 6h
  notifier:
    attempts: 3
    renotifyinterval: 1h
    http:
      # QUAY_ENDPOINT defines the endpoint at which Quay is running.
      # For example: https://myregistry.mycompany.com
      endpoint: { QUAY_ENDPOINT }/secscan/notify
      proxy: http://localhost:6063

jwtproxy:
  signer_proxy:
    enabled: true
    listen_addr: :6063
    ca_key_file: /certificates/mitm.key # Generated internally, do not change.
    ca_crt_file: /certificates/mitm.crt # Generated internally, do not change.
    signer:
      issuer: security_scanner
      expiration_time: 5m
      max_skew: 1m
      nonce_length: 32
      private_key:
        type: autogenerated
        options:
          rotate_every: 12h
          key_folder: /clair/config/
          key_server:
            type: keyregistry
            options:
              # QUAY_ENDPOINT defines the endpoint at which Quay is running.
              # For example: https://myregistry.mycompany.com
              registry: { QUAY_ENDPOINT }/keys/


  verifier_proxies:
  - enabled: true
    # The port at which Clair will listen.
    listen_addr: :6060

    # If Clair is to be served via TLS, uncomment these lines. See the "Running Clair under TLS"
    # section below for more information.
    # key_file: /clair/config/clair.key
    # crt_file: /clair/config/clair.crt

    verifier:
      # CLAIR_ENDPOINT is the endpoint at which this Clair will be accessible. Note that the port
      # specified here must match the listen_addr port a few lines above this.
      # Example: https://myclair.mycompany.com:6060
      audience: { CLAIR_ENDPOINT }

      upstream: http://localhost:6062
      key_server:
        type: keyregistry
        options:
          # QUAY_ENDPOINT defines the endpoint at which Quay is running.
          # Example: https://myregistry.mycompany.com
          registry: { QUAY_ENDPOINT }/keys/

Configuring Clair V2 for TLS

To configure Clair to run with TLS, a few additional steps are required.

Using certificates from a public CA

For certificates that come from a public certificate authority, follow these steps:

  1. Generate a TLS certificate and key pair for the DNS name at which Clair will be accessed

  2. Place these files as clair.crt and clair.key in your Clair configuration directory

  3. Uncomment the key_file and crt_file lines under verifier_proxies in your Clair config.yaml

If your certificates use a public CA, you are now ready to run Clair. If you are using your own certificate authority, configure Clair to trust it below.

Configuring trust of self-signed SSL

Similar to the process for setting up Docker to trust your self-signed certificates, Clair must also be configured to trust your certificates. Using the same CA certificate bundle used to configure Docker, complete the following steps:

  1. Rename the same CA certificate bundle used to set up Quay Registry to ca.crt

  2. Make sure the ca.crt file is mounted inside the Clair container under /etc/pki/ca-trust/source/anchors/ as in the example below: You will need to build your own Clair container and run it during this step. Instructions for building the Clair container are not yet available.

Now Clair will be able to trust the source of your TLS certificates and use them to secure communication between Clair and Quay.

Using Clair V2 data sources

Before scanning container images, Clair tries to figure out the operating system on which the container was built. It does this by looking for specific filenames inside that image (see Table 1). Once Clair knows the operating system, it uses specific security databases to check for vulnerabilities (see Table 2).

Table 2. Container files that identify its operating system
Operating system Files identifying OS type

Redhat/CentOS/Oracle

etc/oracle-release

etc/centos-release

etc/redhat-release

etc/system-release

Alpine

etc/alpine-release

Debian/Ubuntu:

etc/os-release

usr/lib/os-release

etc/apt/sources.list

Ubuntu

etc/lsb-release

The data sources that Clair uses to scan containers are shown in Table 2.

Note

You must be sure that Clair has access to all listed data sources by whitelisting access to each data source’s location. You might need to add a wild-card character (*) at the end of some URLS that may not be fully complete because they are dynamically built by code.

Table 3. Clair V2 data sources and data collected
Data source Data collected Whitelist links Format License

Debian 6, 7, 8, unstable namespaces

Ubuntu 12.04, 12.10, 13.04, 14.04, 14.10, 15.04, 15.10, 16.04 namespaces

CentOS 5, 6, 7 namespace

rpm

Oracle Linux 5, 6, 7 namespaces

rpm

Alpine 3.3, 3.4, 3.5 namespaces

apk

MIT

Generic vulnerability metadata

N/A

Amazon Linux 2018.03, 2 namespaces

rpm

Run Clair V2

Execute the following command to run Clair V2:

You will need to build your own Clair container and run it during this step. Instructions for building the Clair container are not yet available.

Output similar to the following will be seen on success:

2016-05-04 20:01:05,658 CRIT Supervisor running as root (no user in config file)
2016-05-04 20:01:05,662 INFO supervisord started with pid 1
2016-05-04 20:01:06,664 INFO spawned: 'jwtproxy' with pid 8
2016-05-04 20:01:06,666 INFO spawned: 'clair' with pid 9
2016-05-04 20:01:06,669 INFO spawned: 'generate_mitm_ca' with pid 10
time="2016-05-04T20:01:06Z" level=info msg="No claims verifiers specified, upstream should be configured to verify authorization"
time="2016-05-04T20:01:06Z" level=info msg="Starting reverse proxy (Listening on ':6060')"
2016-05-04 20:01:06.715037 I | pgsql: running database migrations
time="2016-05-04T20:01:06Z" level=error msg="Failed to create forward proxy: open /certificates/mitm.crt: no such file or directory"
goose: no migrations to run. current version: 20151222113213
2016-05-04 20:01:06.730291 I | pgsql: database migration ran successfully
2016-05-04 20:01:06.730657 I | notifier: notifier service is disabled
2016-05-04 20:01:06.731110 I | api: starting main API on port 6062.
2016-05-04 20:01:06.736558 I | api: starting health API on port 6061.
2016-05-04 20:01:06.736649 I | updater: updater service is disabled.
2016-05-04 20:01:06,740 INFO exited: jwtproxy (exit status 0; not expected)
2016-05-04 20:01:08,004 INFO spawned: 'jwtproxy' with pid 1278
2016-05-04 20:01:08,004 INFO success: clair entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2016-05-04 20:01:08,004 INFO success: generate_mitm_ca entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
time="2016-05-04T20:01:08Z" level=info msg="No claims verifiers specified, upstream should be configured to verify authorization"
time="2016-05-04T20:01:08Z" level=info msg="Starting reverse proxy (Listening on ':6060')"
time="2016-05-04T20:01:08Z" level=info msg="Starting forward proxy (Listening on ':6063')"
2016-05-04 20:01:08,541 INFO exited: generate_mitm_ca (exit status 0; expected)
2016-05-04 20:01:09,543 INFO success: jwtproxy entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)

To verify Clair V2 is running, execute the following command:

curl -X GET -I http://path/to/clair/here:6061/health

If a 200 OK code is returned, Clair is running:

HTTP/1.1 200 OK
Server: clair
Date: Wed, 04 May 2016 20:02:16 GMT
Content-Length: 0
Content-Type: text/plain; charset=utf-8

Once Clair V2 and its associated database are running, you man need to restart your quay application for the changes to take effect.

Integrating Project Quay into OpenShift Container Platform with the Quay Bridge Operator

Using the Quay Bridge Operator, you can replace the integrated container registry in OpenShift Container Platform with a Project Quay registry. By doing this, your integrated OpenShift Container Platform registry becomes a highly available, enterprise-grade Project Quay registry with enhanced role based access control (RBAC) features.

The primary goal of the Quay Bridge Operator is to duplicate the features of the integrated OpenShift Container Platform registry in the new Project Quay registry. The features enabled with the Quay Bridge Operator include:

  • Synchronizing OpenShift Container Platform namespaces as Project Quay organizations.

  • Creating robot accounts for each default namespace service account.

  • Creating secrets for each created robot account, and associating each robot secret to a service account as Mountable and Image Pull Secret.

  • Synchronizing OpenShift Container Platform image streams as Project Quay repositories.

  • Automatically rewriting new builds making use of image streams to output to Project Quay.

  • Automatically importing an image stream tag once a build completes.

By using the following procedures, you will enable bi-directional communication between your Project Quay and OpenShift Container Platform clusters.

Setting up Project Quay for the Quay Bridge Operator

In this procedure, you will create a dedicated Project Quay organization, and from a new application created within that organization you will generate an OAuth token to be used with the Quay Bridge Operator in OpenShift Container Platform.

Procedure
  1. Log in to Project Quay through the web UI.

  2. Select the organization for which the external application will be configured.

  3. On the navigation pane, select Applications.

  4. Select Create New Application and enter a name for the new application, for example, openshift.

  5. On the OAuth Applications page, select your application, for example, openshift.

  6. On the navigation pane, select Generate Token.

  7. Select the following fields:

    • Administer Organization

    • Administer Repositories

    • Create Repositories

    • View all visible repositories

    • Read/Write to any accessible repositories

    • Administer User

    • Read User Information

  8. Review the assigned permissions.

  9. Select Authorize Application and then confirm confirm the authorization by selecting Authorize Application.

  10. Save the generated access token.

    Important

    As of Project Quay 3.7, there is no token management. You cannot list tokens, delete tokens, or modify tokens. The generated access token is only shown once and cannot be re-obtained after closing the page.

Installing the Quay Bridge Operator on OpenShift Container Platform

In this procedure, you will install the Quay Bridge Operator on OpenShift Container Platform.

Prerequiites
  • You have set up Project Quay and obtained an Access Token.

  • An OpenShift Container Platform 4.6 or greater environment for which you have cluster administrator permissions.

Procedure
  1. Open the Administrator perspective of the web console and navigate to OperatorsOperatorHub on the navigation pane.

  2. Search for Quay Bridge Operator, click the Quay Bridge Operator title, and then click Install.

  3. Select the version to install, for example, stable-3.7, and then click Install.

  4. Click View Operator when the installation finishes to go to the Quay Bridge Operator’s Details page. Alternatively, you can click Installed OperatorsRed Hat Quay Bridge Operator to go to the Details page.

Creating an OpenShift Container Platform secret for the OAuth token

In this procedure, you will add the previously obtained access token to communicate with your Project Quay deployment. The access token will be stored within OpenShift Container Platform as a secret.

Prerequisites
  • You have set up Project Quay and obtained an access token.

  • You have deployed the Quay Bridge Operator on OpenShift Container Platform.

  • An OpenShift Container Platform 4.6 or greater environment for which you have cluster administrator permissions.

  • You have installed the OpenShift CLI (oc).

Procedure
  • Create a secret that contains the access token in the openshift-operators namespace:

    $ oc create secret -n openshift-operators generic <secret-name> --from-literal=token=<access_token>

Creating the QuayIntegration custom resource

In this procedure, you will create a QuayIntegration custom resource, which can be completed from either the web console or from the command line.

Prerequisites
  • You have set up Project Quay and obtained an access token.

  • You have deployed the Quay Bridge Operator on OpenShift Container Platform.

  • An OpenShift Container Platform 4.6 or greater environment for which you have cluster administrator permissions.

  • Optional: You have installed the OpenShift CLI (oc).

Optional: Creating the QuayIntegration custom resource using the CLI

Follow this procedure to create the QuayIntegration custom resource using the command line.

Procedure
  1. Create a quay-integration.yaml:

    $ touch quay-integration.yaml
  2. Use the following configuration for a minimal deployment of the QuayIntegration custom resource:

      apiVersion: quay.redhat.com/v1
      kind: QuayIntegration
      metadata:
        name: example-quayintegration
      spec:
        clusterID: openshift  (1)
        credentialsSecret:
          namespace: openshift-operators
          name: quay-integration(2)
        quayHostname: https://<QUAY_URL>   (3)
        insecureRegistry: false (4)
    1. The clusterID value should be unique across the entire ecosystem. This value is required and defaults to openshift.

    2. The credentialsSecret property refers to the namespace and name of the secret containing the token that was previously created.

    3. Replace the QUAY_URL with the hostname of your Project Quay instance.

    4. If Project Quay is using self signed certificates, set the property to insecureRegistry: true.

    For a list of all configuration fields, see "QuayIntegration configuration fields".

  3. Create the QuayIntegration custom resource:

    $ oc create -f quay-integration.yaml
Optional: Creating the QuayIntegration custom resource using the web console

Follow this procedure to create the QuayIntegration custom resource using the web console.

Procedure
  1. Open the Administrator perspective of the web console and navigate to OperatorsInstalled Operators.

  2. Click Red Hat Quay Bridge Operator.

  3. On the Details page of the Quay Bridge Operator, click Create Instance on the Quay Integration API card.

  4. On the Create QuayIntegration page, enter the following required information in either Form view or YAML view:

    • Name: The name that will refer to the QuayIntegration custom resource object.

    • Cluster ID: The ID associated with this cluster. This value should be unique across the entire ecosystem. Defaults to openshift if left unspecified.

    • Credentials secret: Refers to the namespace and name of the secret containing the token that was previously created.

    • Quay hostname: The hostname of the Quay registry.

      For a list of all configuration fields, see "QuayIntegration configuration fields".

After the QuayIntegration custom resource is created, your OpenShift Container Platform cluster will be linked to your Project Quay instance. Organizations within your Project Quay registry should be created for the related namespace for the OpenShift Container Platform environment.

QuayIntegration configuration fields

The following configuration fields are available for the QuayIntegration custom resource:

Name Description Schema

allowlistNamespaces
(Optional)

A list of namespaces to include.

Array

clusterID
(Required)

The ID associated with this cluster.

String

credentialsSecret.key
(Required)

The secret containing credentials to communicate with the Quay registry.

Object

denylistNamespaces
(Optional)

A list of namespaces to exclude.

Array

insecureRegistry
(Optional)

Whether to skip TLS verification to the Quay registry

Boolean

quayHostname
(Required)

The hostname of the Quay registry.

String

scheduledImageStreamImport
(Optional)

Whether to enable image stream importing.

Boolean

Repository mirroring

Repository mirroring

Project Quay repository mirroring lets you mirror images from external container registries (or another local registry) into your Project Quay cluster. Using repository mirroring, you can synchronize images to Project Quay based on repository names and tags.

From your Project Quay cluster with repository mirroring enabled, you can:

  • Choose a repository from an external registry to mirror

  • Add credentials to access the external registry

  • Identify specific container image repository names and tags to sync

  • Set intervals at which a repository is synced

  • Check the current state of synchronization

To use the mirroring functionality, you need to:

  • Enable repository mirroring in the Project Quay configuration

  • Run a repository mirroring worker

  • Create mirrored repositories

All repository mirroring configuration can be performed using the configuration tool UI or via the Project Quay API

Repository mirroring versus geo-replication

Project Quay geo-replication mirrors the entire image storage backend data between 2 or more different storage backends while the database is shared (one Project Quay registry with two different blob storage endpoints). The primary use cases for geo-replication are:

  • Speeding up access to the binary blobs for geographically dispersed setups

  • Guaranteeing that the image content is the same across regions

Repository mirroring synchronizes selected repositories (or subsets of repositories) from one registry to another. The registries are distinct, with each registry having a separate database and separate image storage. The primary use cases for mirroring are:

  • Independent registry deployments in different datacenters or regions, where a certain subset of the overall content is supposed to be shared across the datacenters / regions

  • Automatic synchronization or mirroring of selected (whitelisted) upstream repositories from external registries into a local Project Quay deployment

Note

Repository mirroring and geo-replication can be used simultaneously.

Table 4. Project Quay Repository mirroring versus geo-replication
Feature / Capability Geo-replication Repository mirroring

What is the feature designed to do?

A shared, global registry

Distinct, different registries

What happens if replication or mirroring hasn’t been completed yet?

The remote copy is used (slower)

No image is served

Is access to all storage backends in both regions required?

Yes (all Project Quay nodes)

No (distinct storage)

Can users push images from both sites to the same repository?

Yes

No

Is all registry content and configuration identical across all regions (shared database)

Yes

No

Can users select individual namespaces or repositories to be mirrored?

No

Yes

Can users apply filters to synchronization rules?

No

Yes

Are individual / different RBAC configurations allowed in each region

No

Yes

Using repository mirroring

Here are some features and limitations of Project Quay repository mirroring:

  • With repository mirroring, you can mirror an entire repository or selectively limit which images are synced. Filters can be based on a comma-separated list of tags, a range of tags, or other means of identifying tags through regular expressions.

  • Once a repository is set as mirrored, you cannot manually add other images to that repository.

  • Because the mirrored repository is based on the repository and tags you set, it will hold only the content represented by the repo / tag pair. In other words, if you change the tag so that some images in the repository no longer match, those images will be deleted.

  • Only the designated robot can push images to a mirrored repository, superseding any role-based access control permissions set on the repository.

  • With a mirrored repository, a user can pull images (given read permission) from the repository but can not push images to the repository.

  • Changing settings on your mirrored repository can be performed in the Project Quay UI, using the Repositories → Mirrors tab for the mirrored repository you create.

  • Images are synced at set intervals, but can also be synced on demand.

Mirroring configuration UI

  1. Start the Quay container in configuration mode and select the Enable Repository Mirroring check box. If you want to require HTTPS communications and verify certificates during mirroring, select the HTTPS and cert verification check box.

    Enable mirroring and require HTTPS and verified certificates

  2. Validate and download the configuration file, and then restart Quay in registry mode using the updated config file.

Mirroring configuration fields

Table 5. Mirroring configuration
Field Type Description

FEATURE_REPO_MIRROR

Boolean

Enable or disable repository mirroring

Default: false

 

 

 

REPO_MIRROR_INTERVAL

Number

The number of seconds between checking for repository mirror candidates

Default: 30

REPO_MIRROR_SERVER_HOSTNAME

String

Replaces the SERVER_HOSTNAME as the destination for mirroring.

Default: None

Example:
openshift-quay-service

REPO_MIRROR_TLS_VERIFY

Boolean

Require HTTPS and verify certificates of Quay registry during mirror.

Default: false

Mirroring worker

  • To run the repository mirroring worker, start by running a Quay pod with the repomirror option:

    $ sudo podman run -d --name mirroring-worker \
      -v $QUAY/config:/conf/stack:Z \
      quay.io/projectquay/quay:v3.7.1 repomirror
  • If you have configured TLS communications using a certificate /root/ca.crt, then the following example shows how to start the mirroring worker:

    $ sudo podman run -d --name mirroring-worker \
      -v $QUAY/config:/conf/stack:Z \
      -v /root/ca.crt:/etc/pki/ca-trust/source/anchors/ca.crt \
      quay.io/projectquay/quay:v3.7.1 repomirror

Creating a mirrored repository

The steps shown in this section assume you already have enabled repository mirroring in the configuration for your Project Quay cluster and that you have a deployed a mirroring worker.

When mirroring a repository from an external container registry, create a new private repository. Typically the same name is used as the target repository, for example, quay-rhel8:

Create new Project Quay repo

Repository mirroring settings

  1. In the Settings tab, set the Repository State to Mirror:

    Create a new Project Quay repo mirror

  2. In the Mirror tab, enter the details for connecting to the external registry, along with the tags, scheduling and access information:

    Repository mirroring

  3. Enter the details as required in the following fields:

    • Registry Location: The external repository you want to mirror, for example, registry.redhat.io/quay/quay-rhel8

    • Tags: This field is required. You may enter a comma-separated list of individual tags or tag patterns. (See Tag Patterns section for details.)

      Note

      In order for Quay to get the list of tags in the remote repository, one of the following requirements must be met:

      • An image with the "latest" tag must exist in the remote repository OR

      • At least one explicit tag, without pattern matching, must exist in the list of tags that you specify

    • Start Date: The date on which mirroring begins. The current date and time is used by default.

    • Sync Interval: Defaults to syncing every 24 hours. You can change that based on hours or days.

    • Robot User: Create a new robot account or choose an existing robot account to do the mirroring.

    • Username: The username for accessing the external registry holding the repository you are mirroring.

    • Password: The password associated with the Username. Note that the password cannot include characters that require an escape character (\).

Advanced settings

  • In the Advanced Settings section, configure TLS and proxy, if required:

  • Verify TLS: Check this box if you want to require HTTPS and to verify certificates, when communicating with the target remote registry.

  • HTTP Proxy: Identify the HTTP proxy server needed to access the remote site, if one is required.

  • HTTPS Proxy: Identify the HTTPS proxy server needed to access the remote site, if one is required.

  • No Proxy: List of locations that do not require proxy

Synchronize now

  • To perform an immediate mirroring operation, press the Sync Now button on the repository’s Mirroring tab. The logs are available on the Usage Logs tab:

    Usage logs

    When the mirroring is complete, the images will appear in the Tags tab:

    Repository mirroring tags

    Below is an example of a completed Repository Mirroring screen:

    Repository mirroring details

Event notifications for mirroring

There are three notification events for repository mirroring:

  • Repository Mirror Started

  • Repository Mirror Success

  • Repository Mirror Unsuccessful

The events can be configured inside the Settings tab for each repository, and all existing notification methods such as email, slack, Quay UI and webhooks are supported.

Mirroring tag patterns

As noted above, at least one Tag must be explicitly entered (ie. not a tag pattern) or the tag "latest" must exist in the report repository. (The tag "latest" will not be synced unless specified in the tag list.). This is required for Quay to get the list of tags in the remote repository to compare to the specified list to mirror.

Pattern syntax

Pattern

Description

*

Matches all characters

?

Matches any single character

[seq]

Matches any character in seq

[!seq]

Matches any character not in seq

Example tag patterns

Example Pattern

Example Matches

v3*

v32, v3.1, v3.2, v3.2-4beta, v3.3

v3.*

v3.1, v3.2, v3.2-4beta

v3.?

v3.1, v3.2, v3.3

v3.[12]

v3.1, v3.2

v3.[12]*

v3.1, v3.2, v3.2-4beta

v3.[!1]*

v3.2, v3.2-4beta, v3.3

Working with mirrored repositories

Once you have created a mirrored repository, there are several ways you can work with that repository. Select your mirrored repository from the Repositories page and do any of the following:

  • Enable/disable the repository: Select the Mirroring button in the left column, then toggle the Enabled check box to enable or disable the repository temporarily.

  • Check mirror logs: To make sure the mirrored repository is working properly, you can check the mirror logs. To do that, select the Usage Logs button in the left column. Here’s an example:

    View logs for your Project Quay repo mirror

  • Sync mirror now: To immediately sync the images in your repository, select the Sync Now button.

  • Change credentials: To change the username and password, select DELETE from the Credentials line. Then select None and add the username and password needed to log into the external registry when prompted.

  • Cancel mirroring: To stop mirroring, which keeps the current images available but stops new ones from being synced, select the CANCEL button.

  • Set robot permissions: Project Quay robot accounts are named tokens that hold credentials for accessing external repositories. By assigning credentials to a robot, that robot can be used across multiple mirrored repositories that need to access the same external registry.

    You can assign an existing robot to a repository by going to Account Settings, then selecting the Robot Accounts icon in the left column. For the robot account, choose the link under the REPOSITORIES column. From the pop-up window, you can:

    • Check which repositories are assigned to that robot.

    • Assign read, write or Admin privileges to that robot from the PERMISSION field shown in this figure: Assign a robot to mirrored repo

  • Change robot credentials: Robots can hold credentials such as Kubernetes secrets, Docker login information, and Mesos bundles. To change robot credentials, select the Options gear on the robot’s account line on the Robot Accounts window and choose View Credentials. Add the appropriate credentials for the external repository the robot needs to access.

    Assign permission to a robot

  • Check and change general setting: Select the Settings button (gear icon) from the left column on the mirrored repository page. On the resulting page, you can change settings associated with the mirrored repository. In particular, you can change User and Robot Permissions, to specify exactly which users and robots can read from or write to the repo.

Repository mirroring recommendations

Best practices for repository mirroring include:

  • Repository mirroring pods can run on any node. This means you can even run mirroring on nodes where Project Quay is already running.

  • Repository mirroring is scheduled in the database and runs in batches. As a result, more workers should mean faster mirroring, since more batches will be processed.

  • The optimal number of mirroring pods depends on:

    • The total number of repositories to be mirrored

    • The number of images and tags in the repositories and the frequency of changes

    • Parallel batches

  • You should balance your mirroring schedule across all mirrored repositories, so that they do not all start up at the same time.

  • For a mid-size deployment, with approximately 1000 users and 1000 repositories, and with roughly 100 mirrored repositories, it is expected that you would use 3-5 mirroring pods, scaling up to 10 pods if required.

LDAP Authentication Setup for Project Quay

The Lightweight Directory Access Protocol (LDAP) is an open, vendor-neutral, industry standard application protocol for accessing and maintaining distributed directory information services over an Internet Protocol (IP) network. Project Quay supports using LDAP as an identity provider.

Considerations prior to enabling LDAP

Existing Quay deployments

Conflicts between user names can arise when you enable LDAP for an existing Quay deployment that already has users configured. Consider the scenario where a particular user, alice, was manually created in Quay prior to enabling LDAP. If the user name alice also exists in the LDAP directory, Quay will create a new user alice-1 when alice logs in for the first time using LDAP, and will map the LDAP credentials to this account. This might not be want you want, for consistency reasons, and it is recommended that you remove any potentially conflicting local account names from Quay prior to enabling LDAP.

Manual User Creation and LDAP authentication

When Quay is configured for LDAP, LDAP-authenticated users are automatically created in Quay’s database on first log in, if the configuration option FEATURE_USER_CREATION is set to true. If this option is set to false, the automatic user creation for LDAP users will fail and the user is not allowed to log in. In this scenario, the superuser needs to create the desired user account first. Conversely, if FEATURE_USER_CREATION is set to true, this also means that a user can still create an account from the Quay login screen, even if there is an equivalent user in LDAP.

Set Up LDAP Configuration

In the config tool, locate the Authentication section and select “LDAP” from the drop-down menu. Update LDAP configuration fields as required.

Fill in LDAP information

  • Here is an example of the resulting entry in the config.yaml file:

AUTHENTICATION_TYPE: LDAP

Full LDAP URI

LDAP server URI LDAP server SSL

  • The full LDAP URI, including the ldap:// or ldaps:// prefix.

  • A URI beginning with ldaps:// will make use of the provided SSL certificate(s) for TLS setup.

  • Here is an example of the resulting entry in the config.yaml file:

LDAP_URI: ldaps://ldap.example.org

Team Synchronization

Team synchronization

  • If enabled, organization administrators who are also superusers can set teams to have their membership synchronized with a backing group in LDAP.

Team synchronization

  • The resynchronization duration is the period at which a team must be re-synchronized. Must be expressed in a duration string form: 30m, 1h, 1d.

  • Optionally allow non-superusers to enable and manage team syncing under organizations in which they are administrators.

  • Here is an example of the resulting entries in the config.yaml file:

FEATURE_TEAM_SYNCING: true
TEAM_RESYNC_STALE_TIME: 60m
FEATURE_NONSUPERUSER_TEAM_SYNCING_SETUP: true

Base and Relative Distinguished Names

Distinguished Names

  • A Distinguished Name path which forms the base path for looking up all LDAP records. Example: dc=my,dc=domain,dc=com

  • Optional list of Distinguished Name path(s) which form the secondary base path(s) for looking up all user LDAP records, relative to the Base DN defined above. These path(s) will be tried if the user is not found via the primary relative DN.

  • User Relative DN is relative to BaseDN. Example: ou=NYC not ou=NYC,dc=example,dc=org

  • Multiple “Secondary User Relative DNs” may be entered if there are multiple Organizational Units where User objects are located at. Simply type in the Organizational Units and click on Add button to add multiple RDNs. Example: ou=Users,ou=NYC and ou=Users,ou=SFO

  • The "User Relative DN" searches with subtree scope. For example, if your Organization has Organizational Units NYC and SFO under the Users OU (ou=SFO,ou=Users and ou=NYC,ou=Users), Project Quay can authenticate users from both the NYC and SFO Organizational Units if the User Relative DN is set to Users (ou=Users).

  • Here is an example of the resulting entries in the config.yaml file:

LDAP_BASE_DN:
- dc=example
- dc=com
LDAP_USER_RDN:
- ou=users
LDAP_SECONDARY_USER_RDNS:
- ou=bots
- ou=external

Additional User Filters

User filters

  • If specified, the additional filter used for all user lookup queries. Note that all Distinguished Names used in the filter must be full paths; the Base DN is not added automatically here. Must be wrapped in parens. Example: (&(someFirstField=someValue)(someOtherField=someOtherValue))

  • Here is an example of the resulting entry in the config.yaml file:

LDAP_USER_FILTER: (memberof=cn=developers,ou=groups,dc=example,dc=com)

Administrator DN

Administrator DN

  • The Distinguished Name and password for the administrator account. This account must be able to login and view the records for all user accounts. Example: uid=admin,ou=employees,dc=my,dc=domain,dc=com

  • The password will be stored in plaintext inside the config.yaml, so setting up a dedicated account or using a password hash is highly recommended.

  • Here is an example of the resulting entries in the config.yaml file:

LDAP_ADMIN_DN: cn=admin,dc=example,dc=com
LDAP_ADMIN_PASSWD: changeme

UID and Mail attributes

UID and Mail

  • The UID attribute is the name of the property field in LDAP user record to use as the username. Typically "uid".

  • The Mail attribute is the name of the property field in LDAP user record that stores user e-mail address(es). Typically "mail".

  • Either of these may be used during login.

  • The logged in username must exist in User Relative DN.

  • sAMAccountName is the UID attribute for against Microsoft Active Directory setups.

  • Here is an example of the resulting entries in the config.yaml file:

LDAP_UID_ATTR: uid
LDAP_EMAIL_ATTR: mail

Validation

Once the configuration is completed, click on “Save Configuration Changes” button to validate the configuration.

Fill in LDAP information

All validation must succeed before proceeding, or additional configuration may be performed by selecting the "Continue Editing" button.

Common Issues

Invalid credentials

Administrator DN or Administrator DN Password values are incorrect

Verification of superuser %USERNAME% failed: Username not found The user either does not exist in the remote authentication system OR LDAP auth is misconfigured.

Project Quay can connect to the LDAP server via Username/Password specified in the Administrator DN fields however cannot find the current logged in user with the UID Attribute or Mail Attribute fields in the User Relative DN Path. Either current logged in user does not exist in User Relative DN Path, or Administrator DN user do not have rights to search/read this LDAP path.

Configure an LDAP user as superuser

Once LDAP is configured, you can log in to your Project Quay instance with a valid LDAP username and password. You are prompted to confirm your Project Quay username as shown in the following figure:

Confirm LDAP username for Project Quay

To attach superuser privilege to an LDAP user, modify the config.yaml file with the username. For example:

SUPER_USERS:
- testadmin

Restart the Red Hat Quay container with the updated config.yaml file. The next time you log in, the user will have superuser privileges.

Prometheus and Grafana metrics under Project Quay

Project Quay exports a Prometheus- and Grafana-compatible endpoint on each instance to allow for easy monitoring and alerting.

Exposing the Prometheus endpoint

Standalone Project Quay

When using podman run to start the Quay container, expose the metrics port 9091:

$ sudo podman run -d --rm -p 80:8080 -p 443:8443  -p 9091:9091\
   --name=quay \
   -v $QUAY/config:/conf/stack:Z \
   -v $QUAY/storage:/datastorage:Z \
   quay.io/projectquay/quay:v3.7.1

The metrics will now be available:

$ curl quay.example.com:9091/metrics

See Monitoring Quay with Prometheus and Grafana for details on configuring Prometheus and Grafana to monitor Quay repository counts.

Project Quay Operator

Determine the cluster IP for the quay-metrics service:

$ oc get services -n quay-enterprise
NAME                                  TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                             AGE
example-registry-clair-app            ClusterIP   172.30.61.161    <none>        80/TCP,8089/TCP                     18h
example-registry-clair-postgres       ClusterIP   172.30.122.136   <none>        5432/TCP                            18h
example-registry-quay-app             ClusterIP   172.30.72.79     <none>        443/TCP,80/TCP,8081/TCP,55443/TCP   18h
example-registry-quay-config-editor   ClusterIP   172.30.185.61    <none>        80/TCP                              18h
example-registry-quay-database        ClusterIP   172.30.114.192   <none>        5432/TCP                            18h
example-registry-quay-metrics         ClusterIP   172.30.37.76     <none>        9091/TCP                            18h
example-registry-quay-redis           ClusterIP   172.30.157.248   <none>        6379/TCP                            18h

Connect to your cluster and access the metrics using the cluster IP and port for the quay-metrics service:

$ oc debug node/master-0

sh-4.4# curl 172.30.37.76:9091/metrics

# HELP go_gc_duration_seconds A summary of the pause duration of garbage collection cycles.
# TYPE go_gc_duration_seconds summary
go_gc_duration_seconds{quantile="0"} 4.0447e-05
go_gc_duration_seconds{quantile="0.25"} 6.2203e-05
...

Setting up Prometheus to consume metrics

Prometheus needs a way to access all Project Quay instances running in a cluster. In the typical setup, this is done by listing all the Project Quay instances in a single named DNS entry, which is then given to Prometheus.

DNS configuration under Kubernetes

A simple Kubernetes service can be configured to provide the DNS entry for Prometheus.

DNS configuration for a manual cluster

SkyDNS is a simple solution for managing this DNS record when not using Kubernetes. SkyDNS can run on an etcd cluster. Entries for each Project Quay instance in the cluster can be added and removed in the etcd store. SkyDNS will regularly read them from there and update the list of Quay instances in the DNS record accordingly.

Introduction to metrics

Project Quay provides metrics to help monitor the registry, including metrics for general registry usage, uploads, downloads, garbage collection, and authentication.

General registry statistics

General registry statistics can indicate how large the registry has grown.

Metric name Description

quay_user_rows

Number of users in the database

quay_robot_rows

Number of robot accounts in the database

quay_org_rows

Number of organizations in the database

quay_repository_rows

Number of repositories in the database

quay_security_scanning_unscanned_images_remaining_total

Number of images that are not scanned by the latest security scanner

Sample metrics output
# HELP quay_user_rows number of users in the database
# TYPE quay_user_rows gauge
quay_user_rows{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="65",process_name="globalpromstats.py"} 3

# HELP quay_robot_rows number of robot accounts in the database
# TYPE quay_robot_rows gauge
quay_robot_rows{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="65",process_name="globalpromstats.py"} 2

# HELP quay_org_rows number of organizations in the database
# TYPE quay_org_rows gauge
quay_org_rows{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="65",process_name="globalpromstats.py"} 2

# HELP quay_repository_rows number of repositories in the database
# TYPE quay_repository_rows gauge
quay_repository_rows{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="65",process_name="globalpromstats.py"} 4

# HELP quay_security_scanning_unscanned_images_remaining number of images that are not scanned by the latest security scanner
# TYPE quay_security_scanning_unscanned_images_remaining gauge
quay_security_scanning_unscanned_images_remaining{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 5

Queue items

The queue items metrics provide information on the multiple queues used by Quay for managing work.

Metric name Description

quay_queue_items_available

Number of items in a specific queue

quay_queue_items_locked

Number of items that are running

quay_queue_items_available_unlocked

Number of items that are waiting to be processed

Metric labels
  • queue_name: The name of the queue. One of:

    • exportactionlogs: Queued requests to export action logs. These logs are then processed and put in storage. A link is then sent to the requester via email.

    • namespacegc: Queued namespaces to be garbage collected

    • notification: Queue for repository notifications to be sent out

    • repositorygc: Queued repositories to be garbage collected

    • secscanv4: Notification queue specific for Clair V4

    • dockerfilebuild: Queue for Quay docker builds

    • imagestoragereplication: Queued blob to be replicated across multiple storages

    • chunk_cleanup: Queued blob segments that needs to be deleted. This is only used by some storage implementations, for example, Swift.

For example, the queue labelled repositorygc contains the repositories marked for deletion by the repository garbage collection worker. For metrics with a queue_name label of repositorygc:

  • quay_queue_items_locked is the number of repositories currently being deleted.

  • quay_queue_items_available_unlocked is the number of repositories waiting to get processed by the worker.

Sample metrics output
# HELP quay_queue_items_available number of queue items that have not expired
# TYPE quay_queue_items_available gauge
quay_queue_items_available{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="63",process_name="exportactionlogsworker.py",queue_name="exportactionlogs"} 0
...

# HELP quay_queue_items_available_unlocked number of queue items that have not expired and are not locked
# TYPE quay_queue_items_available_unlocked gauge
quay_queue_items_available_unlocked{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="63",process_name="exportactionlogsworker.py",queue_name="exportactionlogs"} 0
...

# HELP quay_queue_items_locked number of queue items that have been acquired
# TYPE quay_queue_items_locked gauge
quay_queue_items_locked{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="63",process_name="exportactionlogsworker.py",queue_name="exportactionlogs"} 0

Garbage collection metrics

These metrics show you how many resources have been removed from garbage collection (gc). They show many times the gc workers have run and how many namespaces, repositories, and blobs were removed.

Metric name Description

quay_gc_iterations_total

Number of iterations by the GCWorker

quay_gc_namespaces_purged_total

Number of namespaces purged by the NamespaceGCWorker

quay_gc_repos_purged_total

Number of repositories purged by the RepositoryGCWorker or NamespaceGCWorker

quay_gc_storage_blobs_deleted_total

Number of storage blobs deleted

Sample metrics output
# TYPE quay_gc_iterations_created gauge
quay_gc_iterations_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823190189714e+09
...

# HELP quay_gc_iterations_total number of iterations by the GCWorker
# TYPE quay_gc_iterations_total counter
quay_gc_iterations_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
...

# TYPE quay_gc_namespaces_purged_created gauge
quay_gc_namespaces_purged_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823190189433e+09
...

# HELP quay_gc_namespaces_purged_total number of namespaces purged by the NamespaceGCWorker
# TYPE quay_gc_namespaces_purged_total counter
quay_gc_namespaces_purged_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
....

# TYPE quay_gc_repos_purged_created gauge
quay_gc_repos_purged_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.631782319018925e+09
...

# HELP quay_gc_repos_purged_total number of repositories purged by the RepositoryGCWorker or NamespaceGCWorker
# TYPE quay_gc_repos_purged_total counter
quay_gc_repos_purged_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
...

# TYPE quay_gc_storage_blobs_deleted_created gauge
quay_gc_storage_blobs_deleted_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823190189059e+09
...

# HELP quay_gc_storage_blobs_deleted_total number of storage blobs deleted
# TYPE quay_gc_storage_blobs_deleted_total counter
quay_gc_storage_blobs_deleted_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
...
Multipart uploads metrics

The multipart uploads metrics show the number of blobs uploads to storage (S3, Rados, GoogleCloudStorage, RHOCS). These can help identify issues when Quay is unable to correctly upload blobs to storage.

Metric name Description

quay_multipart_uploads_started_total

Number of multipart uploads to Quay storage that started

quay_multipart_uploads_completed_total

Number of multipart uploads to Quay storage that completed

Sample metrics output
# TYPE quay_multipart_uploads_completed_created gauge
quay_multipart_uploads_completed_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823308284895e+09
...

# HELP quay_multipart_uploads_completed_total number of multipart uploads to Quay storage that completed
# TYPE quay_multipart_uploads_completed_total counter
quay_multipart_uploads_completed_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0

# TYPE quay_multipart_uploads_started_created gauge
quay_multipart_uploads_started_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823308284352e+09
...

# HELP quay_multipart_uploads_started_total number of multipart uploads to Quay storage that started
# TYPE quay_multipart_uploads_started_total counter
quay_multipart_uploads_started_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
...

Image push / pull metrics

A number of metrics are available related to pushing and pulling images.

Image pulls total
Metric name Description

quay_registry_image_pulls_total

The number of images downloaded from the registry.

Metric labels
  • protocol: the registry protocol used (should always be v2)

  • ref: ref used to pull - tag, manifest

  • status: http return code of the request

Image bytes pulled
Metric name Description

quay_registry_image_pulled_estimated_bytes_total

The number of bytes downloaded from the registry

Metric labels
  • protocol: the registry protocol used (should always be v2)

Image pushes total
Metric name Description

quay_registry_image_pushes_total

The number of images uploaded from the registry.

Metric labels
  • protocol: the registry protocol used (should always be v2)

  • pstatus: http return code of the request

  • pmedia_type: the uploaded manifest type

Image bytes pushed
Metric name Description

quay_registry_image_pushed_bytes_total

The number of bytes uploaded to the registry

Sample metrics output
# HELP quay_registry_image_pushed_bytes_total number of bytes pushed to the registry
# TYPE quay_registry_image_pushed_bytes_total counter
quay_registry_image_pushed_bytes_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="221",process_name="registry:application"} 0
...

Authentication metrics

The authentication metrics provide the number of authentication requests, labeled by type and whether it succeeded or not. For example, this metric could be used to monitor failed basic authentication requests.

Metric name Description

quay_authentication_attempts_total

Number of authentication attempts across the registry and API

Metric labels
  • auth_kind: The type of auth used, including:

    • basic

    • oauth

    • credentials

  • success: true or false

Sample metrics output
# TYPE quay_authentication_attempts_created gauge
quay_authentication_attempts_created{auth_kind="basic",host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="221",process_name="registry:application",success="True"} 1.6317843039374158e+09
...

# HELP quay_authentication_attempts_total number of authentication attempts across the registry and API
# TYPE quay_authentication_attempts_total counter
quay_authentication_attempts_total{auth_kind="basic",host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="221",process_name="registry:application",success="True"} 2
...

Project Quay quota management and enforcement

With Project Quay 3.7, users have the ability to report storage consumption and to contain registry growth by establishing configured storage quota limits. On-premise Quay users are now equipped with the following capabilities to manage the capacity limits of their environment:

  • Quota reporting: With this feature, a superuser can track the storage consumption of all their organizations. Additionally, users can track the storage consumption of their assigned organization.

  • Quota management: With this feature, a superuser can define soft and hard checks for Project Quay users. Soft checks tell users if the storage consumption of an organization reaches their configured threshold. Hard checks prevent users from pushing to the registry when storage consumption reaches the configured limit.

Together, these features allow service owners of a Quay registry to define service level agreements and support a healthy resource budget.

Quota management configuration

Quota management is now supported under the FEATURE_QUOTA_MANAGEMENT property and is turned off by default. To enable quota management, set the feature flag in your config.yaml to true:

FEATURE_QUOTA_MANAGEMENT: true
Note

In Project Quay 3.7, superuser privileges are required to create, update and delete quotas. While quotas can be set for users as well as organizations, you cannot reconfigure the user quota using the Project Quay UI and you must use the API instead.

Default quota

To specify a system-wide default storage quota that is applied to every organization and user, use the DEFAULT_SYSTEM_REJECT_QUOTA_BYTES configuration flag.

Table 6. Default quota configuration
Field Type Description

DEFAULT_SYSTEM_REJECT_QUOTA_BYTES

String

The quota size to apply to all organizations and users.

By default, no limit is set.

If you configure a specific quota for an organization or user, and then delete that quota, the system-wide default quota will apply if one has been set. Similarly, if you have configured a specific quota for an organization or user, and then modify the system-wide default quota, the updated system-wide default will override any specific settings.

Quota management architecture

The RepositorySize database table holds the storage consumption, in bytes, of a Project Quay repository within an organization. The sum of all repository sizes for an organization defines the current storage size of a Project Quay organization. When an image push is initialized, the user’s organization storage is validated to check if it is beyond the configured quota limits. If an image push exceeds defined quota limitations, a soft or hard check occurs:

  • For a soft check, users are notified.

  • For a hard check, the push is stopped.

If storage consumption is within configured quota limits, the push is allowed to proceed.

Image manifest deletion follows a similar flow, whereby the links between associated image tags and the manifest are deleted. Additionally, after the image manifest is deleted, the repository size is recalculated and updated in the RepositorySize table.

Establishing quota in Project Quay UI

The following procedure describes how you can report storage consumption and establish storage quota limits.

Prerequisites
  • A Project Quay registry.

  • A superuser account.

  • Enough storage to meet the demands of quota limitations.

Procedure
  1. Create a new organization or choose an existing one. Initially, no quota is configured, as can be seen on the Organization Settings tab:

    No Quota Configured

  2. Log in to the registry as a superuser and navigate to the Manage Organizations tab on the Super User Admin Panel. Click the Options icon of the organization for which you want to create storage quota limits:

    Organization options

  3. Click Configure Quota and enter the initial quota, for example, 10 MB. Then click Apply and Close:

    Initial quota

  4. Check that the quota consumed shows 0 of 10 MB on the Manage Organizations tab of the superuser panel:

    Initial consumed quota

    The consumed quota information is also available directly on the Organization page:

    Initial consumed quota

    Initial consumed quota

  5. To increase the quota to 100MB, navigate to the Manage Organizations tab on the superuser panel. Click the Options icon and select Configure Quota, setting the quota to 100 MB. Click Apply and then Close:

    Increase quota

  6. Push a sample image to the organization from the command line:

    Sample commands
    $ podman pull ubuntu:18.04
    
    $ podman tag docker.io/library/ubuntu:18.04 example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:18.04
    
    $ podman push --tls-verify=false example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:18.04
  7. On the superuser panel, the quota consumed per organization is displayed:

    Total Quota Consumed for first image

  8. The Organization page shows the total proportion of the quota used by the image:

    Total Quota Consumed for first image

    Total Quota Consumed for first image

  9. Pull, tag, and push a second image, for example, nginx:

    Sample commands
    $ podman pull nginx
    
    $ podman tag docker.io/library/nginx example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/nginx
    
    $ podman push --tls-verify=false example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/nginx
  10. The Organization page shows the total proportion of the quota used by each repository in that organization:

    Total Quota Consumed for each repository

    Total Quota Consumed for each repository

  11. Create reject and warning limits:

    From the superuser panel, navigate to the Manage Organizations tab. Click the Options icon for the organization and select Configure Quota. In the Quota Policy section, with the Action type set to Reject, set the Quota Threshold to 80 and click Add Limit:

    Reject limit

  12. To create a warning limit, select Warning as the Action type, set the Quota Threshold to 70 and click Add Limit:

    Warning limit

  13. Click Close on the quota popup. The limits are viewable, but not editable, on the Settings tab of the Organization page:

    Quota policy in organization settings

  14. Push an image where the reject limit is exceeded:

    Because the reject limit (80%) has been set to below the current repository size (~83%), the next push is rejected automatically.

    Sample image push
    $ podman pull ubuntu:20.04
    
    $ podman tag docker.io/library/ubuntu:20.04 example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:20.04
    
    $ podman push --tls-verify=false example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:20.04
    Sample output when quota exceeded
    Getting image source signatures
    Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
    Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
    Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
    WARN[0002] failed, retrying in 1s ... (1/3). Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
    Getting image source signatures
    Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
    Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
    Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
    WARN[0005] failed, retrying in 1s ... (2/3). Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
    Getting image source signatures
    Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
    Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
    Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
    WARN[0009] failed, retrying in 1s ... (3/3). Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
    Getting image source signatures
    Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
    Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
    Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
    Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
  15. When limits are exceeded, notifications are displayed in the UI:

    Quota notifications

    Quota notifications

Establishing quota with the Project Quay API

When an organization is first created, it does not have a quota applied. Use the /api/v1/organization/{organization}/quota endpoint:

Sample command
$ curl -k -X GET -H "Authorization: Bearer <token>" -H 'Content-Type: application/json'  https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/organization/testorg/quota  | jq
Sample output
[]

Setting the quota

To set a quota for an organization, POST data to the /api/v1/organization/{orgname}/quota endpoint: .Sample command

$ curl -k -X POST -H "Authorization: Bearer <token>" -H 'Content-Type: application/json' -d '{"limit_bytes": 10485760}'  https://example-registry-quay-quay-enterprise.apps.docs.quayteam.org/api/v1/organization/testorg/quota | jq
Sample output
"Created"

Viewing the quota

To see the applied quota, GET data from the /api/v1/organization/{orgname}/quota endpoint:

Sample command
$ curl -k -X GET -H "Authorization: Bearer <token>" -H 'Content-Type: application/json'  https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/organization/testorg/quota  | jq
Sample output
[
  {
    "id": 1,
    "limit_bytes": 10485760,
    "default_config": false,
    "limits": [],
    "default_config_exists": false
  }
]

Modifying the quota

To change the existing quota, in this instance from 10 MB to 100 MB, PUT data to the /api/v1/organization/{orgname}/quota/{quota_id} endpoint:

Sample command
$ curl -k -X PUT -H "Authorization: Bearer <token>" -H 'Content-Type: application/json' -d '{"limit_bytes": 104857600}'  https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/organization/testorg/quota/1 | jq
Sample output
{
  "id": 1,
  "limit_bytes": 104857600,
  "default_config": false,
  "limits": [],
  "default_config_exists": false
}

Pushing images

To see the storage consumed, push various images to the organization.

Pushing ubuntu:18.04

Push ubuntu:18.04 to the organization from the command line:

Sample commands
$ podman pull ubuntu:18.04

$ podman tag docker.io/library/ubuntu:18.04 example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:18.04

$ podman push --tls-verify=false example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:18.04
Using the API to view quota usage

To view the storage consumed, GET data from the /api/v1/repository endpoint:

Sample command
$ curl -k -X GET -H "Authorization: Bearer <token>" -H 'Content-Type: application/json' 'https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/repository?last_modified=true&namespace=testorg&popularity=true&public=true&quota=true' | jq
Sample output
{
  "repositories": [
    {
      "namespace": "testorg",
      "name": "ubuntu",
      "description": null,
      "is_public": false,
      "kind": "image",
      "state": "NORMAL",
      "quota_report": {
        "quota_bytes": 27959066,
        "configured_quota": 104857600
      },
      "last_modified": 1651225630,
      "popularity": 0,
      "is_starred": false
    }
  ]
}
Pushing another image
  1. Pull, tag, and push a second image, for example, nginx:

    Sample commands
    $ podman pull nginx
    
    $ podman tag docker.io/library/nginx example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/nginx
    
    $ podman push --tls-verify=false example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/nginx
  2. To view the quota report for the repositories in the organization, use the /api/v1/repository endpoint:

    Sample command
    $ curl -k -X GET -H "Authorization: Bearer <token>" -H 'Content-Type: application/json' 'https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/repository?last_modified=true&namespace=testorg&popularity=true&public=true&quota=true'
    Sample output
    {
      "repositories": [
        {
          "namespace": "testorg",
          "name": "ubuntu",
          "description": null,
          "is_public": false,
          "kind": "image",
          "state": "NORMAL",
          "quota_report": {
            "quota_bytes": 27959066,
            "configured_quota": 104857600
          },
          "last_modified": 1651225630,
          "popularity": 0,
          "is_starred": false
        },
        {
          "namespace": "testorg",
          "name": "nginx",
          "description": null,
          "is_public": false,
          "kind": "image",
          "state": "NORMAL",
          "quota_report": {
            "quota_bytes": 59231659,
            "configured_quota": 104857600
          },
          "last_modified": 1651229507,
          "popularity": 0,
          "is_starred": false
        }
      ]
    }
  3. To view the quota information in the organization details, use the /api/v1/organization/{orgname} endpoint:

    Sample command
    $ curl -k -X GET -H "Authorization: Bearer <token>" -H 'Content-Type: application/json' 'https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/organization/testorg' | jq
    Sample output
    {
      "name": "testorg",
      ...
      "quotas": [
        {
          "id": 1,
          "limit_bytes": 104857600,
          "limits": []
        }
      ],
      "quota_report": {
        "quota_bytes": 87190725,
        "configured_quota": 104857600
      }
    }

Rejecting pushes using quota limits

If an image push exceeds defined quota limitations, a soft or hard check occurs:

  • For a soft check, or warning, users are notified.

  • For a hard check, or reject, the push is terminated.

Setting reject and warning limits

To set reject and warning limits, POST data to the /api/v1/organization/{orgname}/quota/{quota_id}/limit endpoint:

Sample reject limit command
$ curl -k -X POST -H "Authorization: Bearer <token>" -H 'Content-Type: application/json' -d '{"type":"Reject","threshold_percent":80}'  https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/organization/testorg/quota/1/limit
Sample warning limit command
$ curl -k -X POST -H "Authorization: Bearer <token>" -H 'Content-Type: application/json' -d '{"type":"Warning","threshold_percent":50}'  https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/organization/testorg/quota/1/limit
Viewing reject and warning limits

To view the reject and warning limits, use the /api/v1/organization/{orgname}/quota endpoint:

View quota limits
$  curl -k -X GET -H "Authorization: Bearer <token>" -H 'Content-Type: application/json'  https://example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/api/v1/organization/testorg/quota | jq
Sample output for quota limits
[
  {
    "id": 1,
    "limit_bytes": 104857600,
    "default_config": false,
    "limits": [
      {
        "id": 2,
        "type": "Warning",
        "limit_percent": 50
      },
      {
        "id": 1,
        "type": "Reject",
        "limit_percent": 80
      }
    ],
    "default_config_exists": false
  }
]
Pushing an image when the reject limit is exceeded

In this example, the reject limit (80%) has been set to below the current repository size (~83%), so the next push should automatically be rejected.

Push a sample image to the organization from the command line:

Sample image push
$ podman pull ubuntu:20.04

$ podman tag docker.io/library/ubuntu:20.04 example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:20.04

$ podman push --tls-verify=false example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org/testorg/ubuntu:20.04
Sample output when quota exceeded
Getting image source signatures
Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
WARN[0002] failed, retrying in 1s ... (1/3). Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
Getting image source signatures
Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
WARN[0005] failed, retrying in 1s ... (2/3). Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
Getting image source signatures
Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
WARN[0009] failed, retrying in 1s ... (3/3). Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
Getting image source signatures
Copying blob d4dfaa212623 [--------------------------------------] 8.0b / 3.5KiB
Copying blob cba97cc5811c [--------------------------------------] 8.0b / 15.0KiB
Copying blob 0c78fac124da [--------------------------------------] 8.0b / 71.8MiB
Error: Error writing blob: Error initiating layer upload to /v2/testorg/ubuntu/blobs/uploads/ in example-registry-quay-quay-enterprise.apps.docs.gcp.quaydev.org: denied: Quota has been exceeded on namespace
Notifications for limits exceeded

When limits are exceeded, a notification appears:

Quota notifications

Quota notifications

Quota management limitations

Quota management helps organizations to maintain resource consumption. One limitation of quota management is that calculating resource consumption on push results in the calculation becoming part of the push’s critical path. Without this, usage data might drift.

The maximum storage quota size is dependent on the selected database:

Table 7. Worker count environment variables
Variable Description

Postgres

8388608 TB

MySQL

8388608 TB

SQL Server

16777216 TB

Geo-replication

Geo-replication allows multiple, geographically distributed Project Quay deployments to work as a single registry from the perspective of a client or user. It significantly improves push and pull performance in a globally-distributed Project Quay setup. Image data is asynchronously replicated in the background with transparent failover / redirect for clients.

With Project Quay 3.7, deployments of Project Quay with geo-replication is supported by standalone and Operator deployments.

Geo-replication features

  • When geo-replication is configured, container image pushes will be written to the preferred storage engine for that Project Quay instance (typically the nearest storage backend within the region).

  • After the initial push, image data will be replicated in the background to other storage engines.

  • The list of replication locations is configurable and those can be different storage backends.

  • An image pull will always use the closest available storage engine, to maximize pull performance.

  • If replication hasn’t been completed yet, the pull will use the source storage backend instead.

Geo-replication requirements and constraints

  • A single database, and therefore all metadata and Quay configuration, is shared across all regions.

  • A single Redis cache is shared across the entire Quay setup and needs to accessible by all Quay pods.

  • The exact same configuration should be used across all regions, with exception of the storage backend, which can be configured explicitly using the QUAY_DISTRIBUTED_STORAGE_PREFERENCE environment variable.

  • Geo-Replication requires object storage in each region. It does not work with local storage or NFS.

  • Each region must be able to access every storage engine in each region (requires a network path).

  • Alternatively, the storage proxy option can be used.

  • The entire storage backend (all blobs) is replicated. This is in contrast to repository mirroring, which can be limited to an organization or repository or image.

  • All Quay instances must share the same entrypoint, typically via load balancer.

  • All Quay instances must have the same set of superusers, as they are defined inside the common configuration file.

  • Geo-replication requires your Clair configuration to be set to unmanaged. An unmanaged Clair database allows the Project Quay Operator to work in a geo-replicated environment, where multiple instances of the Operator must communicate with the same database. For more information, see Advanced Clair configuration.

  • Geo-Replication requires SSL/TSL certificates and keys. For more information, see Using SSL to protect connections to Project Quay.

If the above requirements cannot be met, you should instead use two or more distinct Quay deployments and take advantage of repository mirroring functionality.

Geo-replication using standalone Project Quay

Georeplication

In the example shown above, Quay is running standalone in two separate regions, with a common database and a common Redis instance. Localized image storage is provided in each region and image pulls are served from the closest available storage engine. Container image pushes are written to the preferred storage engine for the Quay instance, and will then be replicated, in the background, to the other storage engines.

Note

In the event that Clair fails in one cluster, for example, the US cluster, US users would not see vulnerability reports in Quay for the second cluster (EU). This is because all Clair instances have the same state. When Clair fails, it is usually because of a problem within the cluster.

Enable storage replication - standalone Quay

  1. Scroll down to the section entitled Registry Storage.

  2. Click Enable Storage Replication.

  3. Add each of the storage engines to which data will be replicated. All storage engines to be used must be listed.

  4. If complete replication of all images to all storage engines is required, under each storage engine configuration click Replicate to storage engine by default. This will ensure that all images are replicated to that storage engine. To instead enable per-namespace replication, please contact support.

  5. When you are done, click Save Configuration Changes. Configuration changes will take effect the next time Project Quay restarts.

  6. After adding storage and enabling “Replicate to storage engine by default” for Georeplications, you need to sync existing image data across all storage. To do this, you need to oc exec (or docker/kubectl exec) into the container and run:

    # scl enable python27 bash
    # python -m util.backfillreplication

    This is a one time operation to sync content after adding new storage.

Run Project Quay with storage preferences

  1. Copy the config.yaml to all machines running Project Quay

  2. For each machine in each region, add a QUAY_DISTRIBUTED_STORAGE_PREFERENCE environment variable with the preferred storage engine for the region in which the machine is running.

    For example, for a machine running in Europe with the config directory on the host available from $QUAY/config:

    $ sudo podman run -d --rm -p 80:8080 -p 443:8443  \
       --name=quay \
       -v $QUAY/config:/conf/stack:Z \
       -e QUAY_DISTRIBUTED_STORAGE_PREFERENCE=europestorage \
       quay.io/projectquay/quay:v3.7.1
    Note

    The value of the environment variable specified must match the name of a Location ID as defined in the config panel.

  3. Restart all Project Quay containers

Geo-replication using the Project Quay Operator

Georeplication architecture

In the example shown above, the Project Quay Operator is deployed in two separate regions, with a common database and a common Redis instance. Localized image storage is provided in each region and image pulls are served from the closest available storage engine. Container image pushes are written to the preferred storage engine for the Quay instance, and will then be replicated, in the background, to the other storage engines.

Because the Operator now manages the Clair security scanner and its database separately, geo-replication setups can be leveraged so that they do not manage the Clair database. Instead, an external shared database would be used. Project Quay and Clair support several providers and vendors of PostgreSQL, which can be found in the Project Quay 3.x test matrix. Additionally, the Operator also supports custom Clair configurations that can be injected into the deployment, which allows users to configure Clair with the connection credentials for the external database.

Setting up geo-replication on Openshift

Procedure
  1. Deploy Quay postgres instance:

    1. Login to the database

    2. Create a database for Quay

      CREATE DATABASE quay;
    3. Enable pg_trm extension inside the database

      \c quay;
      CREATE EXTENSION IF NOT EXISTS pg_trgm;
  2. Deploy a Redis instance:

    Note
    • Deploying a Redis instance might be unnecessary if your cloud provider has its own service.

    • Deploying a Redis instance is required if you are leveraging Builders.

    1. Deploy a VM for Redis

    2. Make sure that it is accessible from the clusters where Quay is running

    3. Port 6379/TCP must be open

    4. Run Redis inside the instance

      sudo dnf install -y podman
      podman run -d --name redis -p 6379:6379 redis
  3. Create two object storage backends, one for each cluster

    Ideally one object storage bucket will be close to the 1st cluster (primary) while the other will run closer to the 2nd cluster (secondary).

  4. Deploy the clusters with the same config bundle, using environment variable overrides to select the appropriate storage backend for an individual cluster

  5. Configure a load balancer, to provide a single entry point to the clusters

Configuration

The config.yaml file is shared between clusters, and will contain the details for the common PostgreSQL, Redis and storage backends:

config.yaml
SERVER_HOSTNAME: <georep.quayteam.org or any other name> (1)
DB_CONNECTION_ARGS:
  autorollback: true
  threadlocals: true
DB_URI: postgresql://postgres:password@10.19.0.1:5432/quay (2)
BUILDLOGS_REDIS:
  host: 10.19.0.2
  port: 6379
USER_EVENTS_REDIS:
  host: 10.19.0.2
  port: 6379
DISTRIBUTED_STORAGE_CONFIG:
  usstorage:
    - GoogleCloudStorage
    - access_key: GOOGQGPGVMASAAMQABCDEFG
      bucket_name: georep-test-bucket-0
      secret_key: AYWfEaxX/u84XRA2vUX5C987654321
      storage_path: /quaygcp
  eustorage:
    - GoogleCloudStorage
    - access_key: GOOGQGPGVMASAAMQWERTYUIOP
      bucket_name: georep-test-bucket-1
      secret_key: AYWfEaxX/u84XRA2vUX5Cuj12345678
      storage_path: /quaygcp
DISTRIBUTED_STORAGE_DEFAULT_LOCATIONS:
  - usstorage
  - eustorage
DISTRIBUTED_STORAGE_PREFERENCE:
  - usstorage
  - eustorage
FEATURE_STORAGE_REPLICATION: true
  1. A proper SERVER_HOSTNAME must be used for the route and must match the hostname of the global load balancer.

  2. To retrieve the configuration file for a Clair instance deployed using the OpenShift Operator, see Retrieving the Clair config.

Create the configBundleSecret:

$ oc create secret generic --from-file config.yaml=./config.yaml georep-config-bundle

In each of the clusters, set the configBundleSecret and use the QUAY_DISTRIBUTED_STORAGE_PREFERENCE environmental variable override to configure the appropriate storage for that cluster:

Note

The config.yaml file between both deployments must match. If making a change to one cluster, it must also be changed in the other.

US cluster
apiVersion: quay.redhat.com/v1
kind: QuayRegistry
metadata:
  name: example-registry
  namespace: quay-enterprise
spec:
  configBundleSecret: georep-config-bundle
  components:
    - kind: objectstorage
      managed: false
    - kind: route
      managed: true
    - kind: tls
      managed: false
    - kind: postgres
      managed: false
    - kind: clairpostgres
      managed: false
    - kind: redis
      managed: false
    - kind: quay
      managed: true
      overrides:
        env:
          - name: QUAY_DISTRIBUTED_STORAGE_PREFERENCE
            value: usstorage

+

Note

Because TLS is unmanaged, and the route is managed, you must supply the certificates with either with the config tool or directly in the config bundle. For more information, see Configuring TLS and routes.

European cluster
apiVersion: quay.redhat.com/v1
kind: QuayRegistry
metadata:
  name: example-registry
  namespace: quay-enterprise
spec:
  configBundleSecret: georep-config-bundle
  components:
    - kind: objectstorage
      managed: false
    - kind: route
      managed: true
    - kind: tls
      managed: false
    - kind: postgres
      managed: false
    - kind: clairpostgres
      managed: false
    - kind: redis
      managed: false
    - kind: quay
      managed: true
      overrides:
        env:
          - name: QUAY_DISTRIBUTED_STORAGE_PREFERENCE
            value: eustorage

+

Note

Because TLS is unmanaged, and the route is managed, you must supply the certificates with either with the config tool or directly in the config bundle. For more information, see Configuring TLS and routes.

Mixed storage for geo-replication

Project Quay geo-replication supports the use of different and multiple replication targets, for example, using AWS S3 storage on public cloud and using Ceph storage on-prem. This complicates the key requirement of granting access to all storage backends from all Project Quay pods and cluster nodes. As a result, it is recommended that you:

  • Use a VPN to prevent visibility of the internal storage or

  • Use a token pair that only allows access to the specified bucket used by Quay

This will result in the public cloud instance of Project Quay having access to on-prem storage but the network will be encrypted, protected, and will use ACLs, thereby meeting security requirements.

If you cannot implement these security measures, it may be preferable to deploy two distinct Project Quay registries and to use repository mirroring as an alternative to geo-replication.

Backing up and restoring Project Quay managed by the Project Quay Operator

Use the content within this section to back up and restore Project Quay when managed by the Project Quay Operator on OpenShift Container Platform.

Backing up Project Quay

This procedure describes how to create a backup of Project Quay deployed on OpenShift Container Platform using the Project Quay Operator

Prerequisites
  • A healthy Project Quay deployment on OpenShift Container Platform using the Project Quay Operator (status condition Available is set to true)

  • The components quay, postgres and objectstorage are set to managed: true

  • If the component clair is set to managed: true the component clairpostgres is also set to managed: true (starting with Project Quay Operator v3.7 or later)

Note

If your deployment contains partially unmanaged database or storage components and you are using external services for Postgres or S3-compatible object storage to run your Project Quay deployment, you must refer to the service provider or vendor documentation to create a backup of the data. You can refer to the tools described in this guide as a starting point on how to backup your external Postgres database or object storage.

Project Quay configuration backup

  1. Backup the QuayRegistry custom resource by exporting it:

    $ oc get quayregistry <quay-registry-name> -n <quay-namespace> -o yaml > quay-registry.yaml
  2. Edit the resulting quayregistry.yaml and remove the status section and the following metadata fields:

      metadata.creationTimestamp
      metadata.finalizers
      metadata.generation
      metadata.resourceVersion
      metadata.uid
  3. Backup the managed keys secret:

    Note

    If you are running a version older than Project Quay 3.7.0, this step can be skipped. Some secrets are automatically generated while deploying Quay for the first time. These are stored in a secret called <quay-registry-name>-quay-registry-managed-secret-keys in the namespace of the QuayRegistry resource.

    $ oc get secret -n <quay-namespace> <quay-registry-name>-quay-registry-managed-secret-keys -o yaml > managed-secret-keys.yaml
  4. Edit the the resulting managed-secret-keys.yaml file and remove the entry metadata.ownerReferences. Your managed-secret-keys.yaml file should look similar to the following:

    apiVersion: v1
    kind: Secret
    type: Opaque
    metadata:
      name: <quayname>-quay-registry-managed-secret-keys
      namespace: <quay-namespace>
    data:
      CONFIG_EDITOR_PW: <redacted>
      DATABASE_SECRET_KEY: <redacted>
      DB_ROOT_PW: <redacted>
      DB_URI: <redacted>
      SECRET_KEY: <redacted>
      SECURITY_SCANNER_V4_PSK: <redacted>

    All information under the data property should remain the same.

  5. Backup the current Quay configuration:

    $ oc get secret -n <quay-namespace>  $(oc get quayregistry <quay-registry-name> -n <quay-namespace>  -o jsonpath='{.spec.configBundleSecret}') -o yaml > config-bundle.yaml

Scale down your Project Quay deployment

Important

This step is needed to create a consistent backup of the state of your Project Quay deployment. Do not omit this step, including in setups where Postgres databases and/or S3-compatible object storage are provided by external services (unmanaged by the Operator).

  1. For Operator version 3.7 and newer: Scale down the Project Quay deployment by disabling auto scaling and overriding the replica count for Project Quay, mirror workers, and Clair (if managed). Your QuayRegistry resource should look similar to the following:

    apiVersion: quay.redhat.com/v1
    kind: QuayRegistry
    metadata:
      name: registry
      namespace: ns
    spec:
      components:
        …
        - kind: horizontalpodautoscaler
          managed: false (1)
        - kind: quay
          managed: true
          overrides: (2)
            replicas: 0
        - kind: clair
          managed: true
          overrides:
            replicas: 0
        - kind: mirror
          managed: true
          overrides:
            replicas: 0
        …
    1. Disable auto scaling of Quay, Clair and Mirroring workers

    2. Set the replica count to 0 for components accessing the database and objectstorage

  2. For Operator version 3.6 and earlier: Scale down the Project Quay deployment by scaling down the Project Quay Operator first and then the managed Project Quay resources:

    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-operator-namespace>|awk '/^quay-operator/ {print $1}') -n <quay-operator-namespace>
    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-namespace>|awk '/quay-app/ {print $1}') -n <quay-namespace>
    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-namespace>|awk '/quay-mirror/ {print $1}') -n <quay-namespace>
    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-namespace>|awk '/clair-app/ {print $1}') -n <quay-namespace>
  3. Wait for the registry-quay-app, registry-quay-mirror and registry-clair-app pods (depending on which components you set to be managed by the Project Quay Operator) to disappear. You can check their status by running the following command:

    $ oc get pods -n <quay-namespace>

    Example output:

    $ oc get pod
    
    quay-operator.v3.7.1-6f9d859bd-p5ftc               1/1     Running     0             12m
    quayregistry-clair-postgres-7487f5bd86-xnxpr       1/1     Running     1 (12m ago)   12m
    quayregistry-quay-app-upgrade-xq2v6                0/1     Completed   0             12m
    quayregistry-quay-config-editor-6dfdcfc44f-hlvwm   1/1     Running     0             73s
    quayregistry-quay-database-859d5445ff-cqthr        1/1     Running     0             12m
    quayregistry-quay-redis-84f888776f-hhgms           1/1     Running     0             12m

Project Quay managed database backup

Note

If your Project Quay deployment is configured with external (unmanged) Postgres database(s), refer to your vendor’s documentation on how to create a consistent backup of these databases.

  1. Identify the Quay PostgreSQL pod name:

    $ oc get pod -l quay-component=postgres -n <quay-namespace> -o jsonpath='{.items[0].metadata.name}'

    Example output:

    quayregistry-quay-database-59f54bb7-58xs7
  2. Obtain the Quay database name:

    $ oc -n <quay-namespace> rsh $(oc get pod -l app=quay -o NAME -n <quay-namespace> |head -n 1) cat /conf/stack/config.yaml|awk -F"/" '/^DB_URI/ {print $4}'
    quayregistry-quay-database
  3. Download a backup database:

    $ oc exec quayregistry-quay-database-59f54bb7-58xs7 -- /usr/bin/pg_dump -C quayregistry-quay-database  > backup.sql
Project Quay managed object storage backup

The instructions in this section apply to the following configurations:

  • Standalone, multi-cloud object gateway configurations

  • OpenShift Data Foundations storage requires that the Project Quay Operator provisioned an S3 object storage bucket from, through the ObjectStorageBucketClaim API

Note

If your Project Quay deployment is configured with external (unmanged) object storage, refer to your vendor’s documentation on how to create a copy of the content of Quay’s storage bucket.

  1. Decode and export the AWS_ACCESS_KEY_ID:

    $ export AWS_ACCESS_KEY_ID=$(oc get secret -l app=noobaa -n <quay-namespace>  -o jsonpath='{.items[0].data.AWS_ACCESS_KEY_ID}' |base64 -d)
  2. Decode and export the AWS_SECRET_ACCESS_KEY_ID:

    $ export AWS_SECRET_ACCESS_KEY=$(oc get secret -l app=noobaa -n <quay-namespace> -o jsonpath='{.items[0].data.AWS_SECRET_ACCESS_KEY}' |base64 -d)
  3. Create a new directory and copy all blobs to it:

    $ mkdir blobs
    
    $ aws s3 sync --no-verify-ssl --endpoint https://$(oc get route s3 -n openshift-storage  -o jsonpath='{.spec.host}')  s3://$(oc get cm -l app=noobaa -n <quay-namespace> -o jsonpath='{.items[0].data.BUCKET_NAME}') ./blobs
Note

You can also use rclone or sc3md instead of the AWS command line utility.

Scale the Project Quay deployment back up

  1. For Operator version 3.7 and newer: Scale up the Project Quay deployment by re-enabling auto scaling, if desired, and removing the replica overrides for Quay, mirror workers and Clair as applicable. Your QuayRegistry resource should look similar to the following:

    apiVersion: quay.redhat.com/v1
    kind: QuayRegistry
    metadata:
      name: registry
      namespace: ns
    spec:
      components:
        …
        - kind: horizontalpodautoscaler
          managed: true (1)
        - kind: quay (2)
          managed: true
        - kind: clair
          managed: true
        - kind: mirror
          managed: true
        …
    1. Re-enables auto scaling of Quay, Clair and Mirroring workers again (if desired)

    2. Replica overrides are removed again to scale the Quay components back up

  2. For Operator version 3.6 and earlier: Scale up the Project Quay deployment by scaling up the Project Quay Operator again:

    $ oc scale --replicas=1 deployment $(oc get deployment -n <quay-operator-namespace> | awk '/^quay-operator/ {print $1}') -n <quay-operator-namespace>
  3. Check the status of the Project Quay deployment:

    $ oc wait quayregistry registry --for=condition=Available=true -n <quay-namespace>

    Example output:

    apiVersion: quay.redhat.com/v1
    kind: QuayRegistry
    metadata:
      ...
      name: registry
      namespace: <quay-namespace>
      ...
    spec:
      ...
    status:
      - lastTransitionTime: '2022-06-20T05:31:17Z'
        lastUpdateTime: '2022-06-20T17:31:13Z'
        message: All components reporting as healthy
        reason: HealthChecksPassing
        status: 'True'
        type: Available

Restoring Project Quay

This procedure is used to restore Project Quay when the Project Quay Operator manages the database. It should be performed after a backup of your Project Quay registry has been performed. See Backing up Project Quay for more information.

Prerequisites
  • Project Quay is deployed on OpenShift Container Platform using the Project Quay Operator.

  • A backup of the Project Quay configuration managed by the Project Quay Operator has been created following the instructions in the Backing up Project Quay section

  • Your Project Quay database has been backed up.

  • The object storage bucket used by Project Quay has been backed up.

  • The components quay, postgres and objectstorage are set to managed: true

  • If the component clair is set to managed: true, the component clairpostgres is also set to managed: true (starting with Project Quay Operator v3.7 or later)

  • There is no running Project Quay deployment managed by the Project Quay Operator in the target namespace on your OpenShift Container Platform cluster

Note

If your deployment contains partially unmanaged database or storage components and you are using external services for Postgres or S3-compatible object storage to run your Project Quay deployment, you must refer to the service provider or vendor documentation to restore their data from a backup prior to restore Project Quay

Restoring Project Quay and its configuration from a backup

Note

These instructions assume you have followed the process in the Backing up Project Quay guide and create the backup files with the same names.

  1. Restore the backed up Project Quay configuration and the generated keys from the backup:

    $ oc create -f ./config-bundle.yaml
    
    $ oc create -f ./managed-secret-keys.yaml
    Important

    If you receive the error Error from server (AlreadyExists): error when creating "./config-bundle.yaml": secrets "config-bundle-secret" already exists, you must delete your existing resource with $ oc delete Secret config-bundle-secret -n <quay-namespace> and recreate it with $ oc create -f ./config-bundle.yaml.

  2. Restore the QuayRegistry custom resource:

    $ oc create -f ./quay-registry.yaml
  3. Check the status of the Project Quay deployment and wait for it to be available:

    $ oc wait quayregistry registry --for=condition=Available=true -n <quay-namespace>

Scale down your Project Quay deployment

  1. For Operator version 3.7 and newer: Scale down the Project Quay deployment by disabling auto scaling and overriding the replica count for Quay, mirror workers and Clair (if managed). Your QuayRegistry resource should look similar to the following:

    apiVersion: quay.redhat.com/v1
    kind: QuayRegistry
    metadata:
      name: registry
      namespace: ns
    spec:
      components:
        …
        - kind: horizontalpodautoscaler
          managed: false (1)
        - kind: quay
          managed: true
          overrides: (2)
            replicas: 0
        - kind: clair
          managed: true
          overrides:
            replicas: 0
        - kind: mirror
          managed: true
          overrides:
            replicas: 0
        …
    1. Disable auto scaling of Quay, Clair and Mirroring workers

    2. Set the replica count to 0 for components accessing the database and objectstorage

  2. For Operator version 3.6 and earlier: Scale down the Project Quay deployment by scaling down the Project Quay Operator first and then the managed Project Quay resources:

    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-operator-namespace>|awk '/^quay-operator/ {print $1}') -n <quay-operator-namespace>
    
    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-namespace>|awk '/quay-app/ {print $1}') -n <quay-namespace>
    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-namespace>|awk '/quay-mirror/ {print $1}') -n <quay-namespace>
    $ oc scale --replicas=0 deployment $(oc get deployment -n <quay-namespace>|awk '/clair-app/ {print $1}') -n <quay-namespace>
  3. Wait for the registry-quay-app, registry-quay-mirror and registry-clair-app pods (depending on which components you set to be managed by Operator) to disappear. You can check their status by running the following command:

    $ oc get pods -n <quay-namespace>

    Example output:

    registry-quay-config-editor-77847fc4f5-nsbbv   1/1     Running            0          9m1s
    registry-quay-database-66969cd859-n2ssm        1/1     Running            0          6d1h
    registry-quay-redis-7cc5f6c977-956g8           1/1     Running            0          5d21h

Restore your Project Quay database

  1. Identify your Quay database pod:

    $ oc get pod -l quay-component=postgres -n  <quay-namespace> -o jsonpath='{.items[0].metadata.name}'

    Example output:

    quayregistry-quay-database-59f54bb7-58xs7
  2. Upload the backup by copying it from the local environment and into the pod:

    $ oc cp ./backup.sql -n <quay-namespace> registry-quay-database-66969cd859-n2ssm:/tmp/backup.sql
  3. Open a remote terminal to the database:

    $ oc rsh -n <quay-namespace> registry-quay-database-66969cd859-n2ssm
  4. Enter psql:

    bash-4.4$ psql
  5. You can list the database by running the following command:

    postgres=# \l

    Example output:

                                                      List of databases
               Name            |           Owner            | Encoding |  Collate   |   Ctype    |   Access privileges
    ----------------------------+----------------------------+----------+------------+------------+-----------------------
    postgres                   | postgres                   | UTF8     | en_US.utf8 | en_US.utf8 |
    quayregistry-quay-database | quayregistry-quay-database | UTF8     | en_US.utf8 | en_US.utf8 |
  6. Drop the database:

    postgres=# DROP DATABASE "quayregistry-quay-database";

    Example output:

    DROP DATABASE
  7. Exit the postgres CLI to re-enter bash-4.4:

    \q
  8. Redirect your PostgreSQL database to your backup database:

    sh-4.4$ psql < /tmp/backup.sql
  9. Exit bash:

    sh-4.4$ exit

Restore your Project Quay object storage data

  1. Export the AWS_ACCESS_KEY_ID:

    $ export AWS_ACCESS_KEY_ID=$(oc get secret -l app=noobaa -n <quay-namespace>  -o jsonpath='{.items[0].data.AWS_ACCESS_KEY_ID}' |base64 -d)
  2. Export the AWS_SECRET_ACCESS_KEY:

    $ export AWS_SECRET_ACCESS_KEY=$(oc get secret -l app=noobaa -n <quay-namespace> -o jsonpath='{.items[0].data.AWS_SECRET_ACCESS_KEY}' |base64 -d)
  3. Upload all blobs to the bucket by running the following command:

    $ aws s3 sync --no-verify-ssl --endpoint https://$(oc get route s3 -n openshift-storage  -o jsonpath='{.spec.host}') ./blobs  s3://$(oc get cm -l app=noobaa -n <quay-namespace> -o jsonpath='{.items[0].data.BUCKET_NAME}')
Note

You can also use rclone or sc3md instead of the AWS command line utility.

Scale up your Project Quay deployment

  1. For Operator version 3.7 and newer: Scale up the Project Quay deployment by re-enabling auto scaling, if desired, and removing the replica overrides for Quay, mirror workers and Clair as applicable. Your QuayRegistry resource should look similar to the following:

    apiVersion: quay.redhat.com/v1
    kind: QuayRegistry
    metadata:
      name: registry
      namespace: ns
    spec:
      components:
        …
        - kind: horizontalpodautoscaler
          managed: true (1)
        - kind: quay (2)
          managed: true
        - kind: clair
          managed: true
        - kind: mirror
          managed: true
        …
    1. Re-enables auto scaling of Project Quay, Clair and mirroring workers again (if desired)

    2. Replica overrides are removed again to scale the Project Quay components back up

  2. For Operator version 3.6 and earlier: Scale up the Project Quay deployment by scaling up the Project Quay Operator again:

    $ oc scale --replicas=1 deployment $(oc get deployment -n <quay-operator-namespace> | awk '/^quay-operator/ {print $1}') -n <quay-operator-namespace>
  3. Check the status of the Project Quay deployment:

    $ oc wait quayregistry registry --for=condition=Available=true -n <quay-namespace>

    Example output:

    apiVersion: quay.redhat.com/v1
    kind: QuayRegistry
    metadata:
      ...
      name: registry
      namespace: <quay-namespace>
      ...
    spec:
      ...
    status:
      - lastTransitionTime: '2022-06-20T05:31:17Z'
        lastUpdateTime: '2022-06-20T17:31:13Z'
        message: All components reporting as healthy
        reason: HealthChecksPassing
        status: 'True'
        type: Available

Migrating a standalone Quay deployment to a Project Quay Operator managed deployment

The following procedures allow you to back up a standalone Project Quay deployment and migrate it to the Project Quay Operator on OpenShift Container Platform.

Backing up a standalone deployment of Project Quay

Procedure
  1. Back up the Quay config.yaml of your standalone deployment:

    $ mkdir /tmp/quay-backup
    $ cp /path/to/Quay/config/directory/config.yaml /tmp/quay-backup
  2. Create a backup of the database that your standalone Quay deployment is using:

    $ pg_dump -h DB_HOST -p 5432 -d QUAY_DATABASE_NAME -U QUAY_DATABASE_USER -W -O > /tmp/quay-backup/quay-database-backup.sql
  3. Install the AWS CLI if you do not have it already.

  4. Create an ~/.aws/ directory:

    $ mkdir ~/.aws/
  5. Obtain the access_key and secret_key from the Quay config.yaml of your standalone deployment:

    $ grep -i DISTRIBUTED_STORAGE_CONFIG -A10 /tmp/quay-backup/config.yaml

    Example output:

    DISTRIBUTED_STORAGE_CONFIG:
        minio-1:
            - RadosGWStorage
            - access_key: ##########
              bucket_name: quay
              hostname: 172.24.10.50
              is_secure: false
              port: "9000"
              secret_key: ##########
              storage_path: /datastorage/registry
  6. Store the access_key and secret_key from the Quay config.yaml file in your ~/.aws directory:

    $ touch ~/.aws/credentials
  7. Optional: Check that your access_key and secret_key are stored:

    $ cat > ~/.aws/credentials << EOF
    [default]
    aws_access_key_id = ACCESS_KEY_FROM_QUAY_CONFIG
    aws_secret_access_key = SECRET_KEY_FROM_QUAY_CONFIG
    EOF

    Example output:

    aws_access_key_id = ACCESS_KEY_FROM_QUAY_CONFIG
    aws_secret_access_key = SECRET_KEY_FROM_QUAY_CONFIG
    Note

    If the aws cli does not automatically collect the access_key and secret_key from the `~/.aws/credentials file, you can, you can configure these by running aws configure and manually inputting the credentials.

  8. In your quay-backup directory, create a bucket_backup directory:

    $ mkdir /tmp/quay-backup/bucket-backup
  9. Backup all blobs from the S3 storage:

    $ aws s3 sync --no-verify-ssl --endpoint-url https://PUBLIC_S3_ENDPOINT:PORT s3://QUAY_BUCKET/ /tmp/quay-backup/bucket-backup/
    Note

    The PUBLIC_S3_ENDPOINT can be read from the Quay config.yaml file under hostname in the DISTRIBUTED_STORAGE_CONFIG. If the endpoint is insecure, use http instead of https in the endpoint URL.

Up to this point, you should have a complete backup of all Quay data, blobs, the database, and the config.yaml file stored locally. In the following section, you will migrate the standalone deployment backup to Project Quay on OpenShift Container Platform.

Using backed up standalone content to migrate to OpenShift Container Platform.

Prerequisites
  • Your standalone Project Quay data, blobs, database, and config.yaml have been backed up.

  • Project Quay is deployed on OpenShift Container Platform using the Quay Operator.

  • A QuayRegistry with all components set to managed.

Note
Procedure

The procedure in this documents uses the following namespace: quay-enterprise.

  1. Scale down the Project Quay Operator:

    $ oc scale --replicas=0 deployment quay-operator.v3.6.2 -n openshift-operators
  2. Scale down the application and mirror deployments:

    $ oc scale --replicas=0 deployment QUAY_MAIN_APP_DEPLOYMENT QUAY_MIRROR_DEPLOYMENT
  3. Copy the database SQL backup to the Quay PostgreSQL database instance:

    $ oc cp /tmp/user/quay-backup/quay-database-backup.sql quay-enterprise/quayregistry-quay-database-54956cdd54-p7b2w:/var/lib/pgsql/data/userdata
  4. Obtain the database password from the Operator-created config.yaml file:

    $ oc get deployment quay-quay-app -o json | jq '.spec.template.spec.volumes[].projected.sources' | grep -i config-secret

    Example output:

          "name": "QUAY_CONFIG_SECRET_NAME"
    $ oc get secret quay-quay-config-secret-9t77hb84tb -o json | jq '.data."config.yaml"' | cut -d '"' -f2 | base64 -d -w0 > /tmp/quay-backup/operator-quay-config-yaml-backup.yaml
    cat /tmp/quay-backup/operator-quay-config-yaml-backup.yaml | grep -i DB_URI

    Example output:

    postgresql://QUAY_DATABASE_OWNER:PASSWORD@DATABASE_HOST/QUAY_DATABASE_NAME
  5. Execute a shell inside of the database pod:

    # oc exec -it quay-postgresql-database-pod -- /bin/bash
  6. Enter psql:

    bash-4.4$ psql
  7. Drop the database:

    postgres=# DROP DATABASE "example-restore-registry-quay-database";

    Example output:

    DROP DATABASE
  8. Create a new database and set the owner as the same name:

    postgres=# CREATE DATABASE "example-restore-registry-quay-database" OWNER "example-restore-registry-quay-database";

    Example output:

    CREATE DATABASE
  9. Connect to the database:

    postgres=# \c "example-restore-registry-quay-database";

    Example output:

    You are now connected to database "example-restore-registry-quay-database" as user "postgres".
  10. Create a pg_trmg extension of your Quay database:

    example-restore-registry-quay-database=# create extension pg_trgm ;

    Example output:

    CREATE EXTENSION
  11. Exit the postgres CLI to re-enter bash-4.4:

    \q
  12. Set the password for your PostgreSQL deployment:

    bash-4.4$ psql -h localhost -d "QUAY_DATABASE_NAME" -U QUAY_DATABASE_OWNER -W < /var/lib/pgsql/data/userdata/quay-database-backup.sql

    Example output:

    SET
    SET
    SET
    SET
    SET
  13. Exit bash mode:

    bash-4.4$ exit
  14. Create a new configuration bundle for the Project Quay Operator.

    $ touch config-bundle.yaml
  15. In your new config-bundle.yaml, include all of the information that the registry requires, such as LDAP configuration, keys, and other modifications that your old registry had. Run the following command to move the secret_key to your config-bundle.yaml:

    $ cat /tmp/quay-backup/config.yaml | grep SECRET_KEY > /tmp/quay-backup/config-bundle.yaml
    Note

    You must manually copy all the LDAP, OIDC and other information and add it to the /tmp/quay-backup/config-bundle.yaml file.

  16. Create a configuration bundle secret inside of your OpenShift cluster:

    $ oc create secret generic new-custom-config-bundle --from-file=config.yaml=/tmp/quay-backup/config-bundle.yaml
  17. Scale up the Quay pods:

    $ oc scale --replicas=1 deployment quayregistry-quay-app
    deployment.apps/quayregistry-quay-app scaled
  18. Scale up the mirror pods:

    $ oc scale --replicas=1  deployment quayregistry-quay-mirror
    deployment.apps/quayregistry-quay-mirror scaled
  19. Patch the QuayRegistry CRD so that it contains the reference to the new custom configuration bundle:

    $ oc patch quayregistry QUAY_REGISTRY_NAME --type=merge -p '{"spec":{"configBundleSecret":"new-custom-config-bundle"}}'
    Note

    If Quay returns a 500 internal server error, you might have to update the location of your DISTRIBUTED_STORAGE_CONFIG to default.

  20. Create a new AWS credentials.yaml in your /.aws/ directory and include the access_key and secret_key from the Operator-created config.yaml file:

    $ touch credentials.yaml
    $ grep -i DISTRIBUTED_STORAGE_CONFIG -A10 /tmp/quay-backup/operator-quay-config-yaml-backup.yaml
    $ cat > ~/.aws/credentials << EOF
    [default]
    aws_access_key_id = ACCESS_KEY_FROM_QUAY_CONFIG
    aws_secret_access_key = SECRET_KEY_FROM_QUAY_CONFIG
    EOF
    Note

    If the aws cli does not automatically collect the access_key and secret_key from the `~/.aws/credentials file, you can configure these by running aws configure and manually inputting the credentials.

  21. Record the NooBaa’s publicly available endpoint:

    $ oc get route s3 -n openshift-storage -o yaml -o jsonpath="{.spec.host}{'\n'}"
  22. Sync the backup data to the NooBaa backend storage:

    $ aws s3 sync --no-verify-ssl --endpoint-url https://NOOBAA_PUBLIC_S3_ROUTE /tmp/quay-backup/bucket-backup/* s3://QUAY_DATASTORE_BUCKET_NAME
  23. Scale the Operator back up to 1 pod:

    $ oc scale –replicas=1 deployment quay-operator.v3.6.4 -n openshift-operators

The Operator will use the custom configuration bundle provided and will reconcile all secrets and deployments. Your new Quay deployment on OpenShift Container Platform should contain all of the information that the old deployment had. All images should be pull-able.

Project Quay garbage collection

About Project Quay garbage collection

Project Quay includes automatic and continuous image garbage collection. Garbage collection ensures efficient use of resources for active objects by removing objects that occupy sizeable amounts of disk space, such as dangling or untagged images, repositories, and blobs, including layers and manifests. Garbage collection performed by Project Quay can reduce downtime in your organization’s environment.

Project Quay garbage collection in practice

Currently, all garbage collection happens discreetly; there are no commands to manually run garbage collection. Project Quay provides metrics that track the status of the different garbage collection workers.

For namespace and repository garbage collection, the progress is tracked based on the size of their respective queues. Namespace and repository garbage collection workers require a global lock to work. As a result, and for performance reasons, only one worker runs at a time.

Garbage collecting tagged images works differently than garbage collection on namespaces or repositories. Rather than having a queue of items to work with, the garbage collection workers for tagged images actively search for a repository with inactive or expired tags to clean up. Each instance of garbage collection workers will grab a repository lock, which results in one worker per repository.

For each type of garbage collection, Project Quay provides metrics for the number of rows per table deleted by each garbage collection worker. The following image shows an example of how Project Quay monitors garbage collection with the same metrics:

Garbage collection metrics

Measuring storage reclamation

Project Quay does not have a way to track how much space is freed up by garbage collection. Currently, the best indicator of this is by checking how many blobs have been deleted in the provided metrics.

Note

The UploadedBlob table in the Project Quay metrics tracks the various blobs that are associated with a repository. When a blob is uploaded, it will not be garbage collected before the time designated by the PUSH_TEMP_TAG_EXPIRATION_SEC parameter. This is to avoid prematurely deleting blobs that are part of an ongoing push. For example, if garbage collection is set to run often, and a tag is deleted in the span of less than one hour, then it is possible that the associated blobs will not get cleaned up immediately. Instead, and assuming that the time designated by the PUSH_TEMP_TAG_EXPIRATION_SEC parameter has passed, the associated blobs will be removed the next time garbage collection runs on that same repository.

Garbage collection configuration fields

The following configuration fields are available to customize what is garbage collected, and the frequency at which garbage collection occurs:

Name Description Schema

FEATURE_GARBAGE_COLLECTION

Whether garbage collection is enabled for image tags. Defaults to true.

Boolean

FEATURE_NAMESPACE_GARBAGE_COLLECTION

Whether garbage collection is enabled for namespaces. Defaults to true.

Boolean

FEATURE_REPOSITORY_GARBAGE_COLLECTION

Whether garbage collection is enabled for repositories. Defaults to true.

Boolean

GARBAGE_COLLECTION_FREQUENCY

The frequency, in seconds, at which the garbage collection worker runs. Affects only garbage collection workers. Defaults to 30 seconds.

String

PUSH_TEMP_TAG_EXPIRATION_SEC

The number of seconds that blobs will not be garbage collected after being uploaded. This feature prevents garbage collection from cleaning up blobs that are not referenced yet, but still used as part of an ongoing push.

String

TAG_EXPIRATION_OPTIONS

List of valid tag expiration values.

String

DEFAULT_TAG_EXPIRATION

Tag expiration time for time machine.

String

Disabling garbage collection

The garbage collection features for image tags, namespaces, and repositories are stored in the config.yaml file. These features default to true.

In rare cases, you might want to disable garbage collection, for example, to control when garbage collection is performed. You can disable garbage collection by setting the GARBAGE_COLLECTION features to false. When disabled, dangling or untagged images, repositories, namespaces, layers, and manifests are not removed. This might increase the downtime of your environment.

Note

There is no command to manually run garbage collection. Instead, you would disable, and then re-enable, the garbage collection feature.

Garbage collection and quota management

Project Quay introduced quota management in 3.7. With quota management, users have the ability to report storage consumption and to contain registry growth by establishing configured storage quota limits.

As of Project Quay 3.7, garbage collection reclaims memory that was allocated to images, repositories, and blobs after deletion. Because the garbage collection feature reclaims memory after deletion, there is a discrepancy between what is stored in an environment’s disk space and what quota management is reporting as the total consumption. There is currently no workaround for this issue.

Project Quay garbage collection metrics

The following metrics show how many resources have been removed by garbage collection. These metrics show how many times the garbage collection workers have run and how many namespaces, repositories, and blobs were removed.

Metric name Description

quay_gc_iterations_total

Number of iterations by the GCWorker

quay_gc_namespaces_purged_total

Number of namespaces purged by the NamespaceGCWorker

quay_gc_repos_purged_total

Number of repositories purged by the RepositoryGCWorker or NamespaceGCWorker

quay_gc_storage_blobs_deleted_total

Number of storage blobs deleted

Sample metrics output
# TYPE quay_gc_iterations_created gauge
quay_gc_iterations_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823190189714e+09
...

# HELP quay_gc_iterations_total number of iterations by the GCWorker
# TYPE quay_gc_iterations_total counter
quay_gc_iterations_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
...

# TYPE quay_gc_namespaces_purged_created gauge
quay_gc_namespaces_purged_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823190189433e+09
...

# HELP quay_gc_namespaces_purged_total number of namespaces purged by the NamespaceGCWorker
# TYPE quay_gc_namespaces_purged_total counter
quay_gc_namespaces_purged_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
....

# TYPE quay_gc_repos_purged_created gauge
quay_gc_repos_purged_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.631782319018925e+09
...

# HELP quay_gc_repos_purged_total number of repositories purged by the RepositoryGCWorker or NamespaceGCWorker
# TYPE quay_gc_repos_purged_total counter
quay_gc_repos_purged_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
...

# TYPE quay_gc_storage_blobs_deleted_created gauge
quay_gc_storage_blobs_deleted_created{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 1.6317823190189059e+09
...

# HELP quay_gc_storage_blobs_deleted_total number of storage blobs deleted
# TYPE quay_gc_storage_blobs_deleted_total counter
quay_gc_storage_blobs_deleted_total{host="example-registry-quay-app-6df87f7b66-9tfn6",instance="",job="quay",pid="208",process_name="secscan:application"} 0
...

Project Quay Troubleshooting

Schema for Project Quay configuration

Most Project Quay configuration information is stored in the config.yaml file that is created using the browser-based config tool when Project Quay is first deployed.

The configuration options are described in the Project Quay Configuration Guide.

Additional resources