Chapter 3. Installing Red Hat Quay (high availability)

Note

This procedure presents guidance on how to set up a highly available, production-quality deployment of Red Hat Quay.

3.1. Prerequisites

Here are are a few things you need to know before you begin the Red Hat Quay high availability deployment:

  • Either Postgres or MySQL can be used to provide the database service. Postgres was chosen here as the database because it includes the features needed to support Clair security scanning.
Note

You can substitute your own enterprise-quality database if you choose. The Postgres database illustrated here is not, itself, configured for high availability.

  • Ceph Object Gateway (also called RADOS Gateway) provides the object storage needed by Red Hat Quay. If you want your Red Hat Quay setup to do geo-replication, Ceph Object Gateway or other supported object storage is required. For cloud installations, you can use any of the following cloud object storage:

    • Amazon S3 (see S3 IAM Bucket Policy for details on configuring an S3 bucket policy for Quay)
    • Azure Blob Storage
    • Google Cloud Storage
    • Ceph Object Gateway
    • OpenStack Swift
    • CloudFront + S3
  • The haproxy server is used in this example, although you can use any proxy service that works for your environment.
  • Number of systems: This procedure uses seven systems (physical or virtual) that are assigned with the following tasks:

    • db01: Load balancer and database: Runs the haproxy load balancer and a Postgres database. Note that thes components are not themselves highly available, but are used to indicate how you might set up your own load balancer or production database.
    • quay01, quay02, quay03: Quay and Redis: Three (or more) systems are assigned to run the Quay and Redis services.
    • ceph01, ceph02, ceph03, ceph04, ceph05: Ceph: Three (or more) systems provide the Ceph service, for storage. If you are deploying to a cloud, you can use the cloud storage features described earlier. This procedure employs an additional system for Ansible (ceph05) and one for a Ceph Object Gateway (ceph04).

Each system should have the following attributes:

  • Red Hat Enterprise Linux (RHEL): Obtain the latest Red Hat Enterprise Linux server media from the Downloads page and follow instructions from the Red Hat Enterprise Linux 7 Installation Guide to install RHEL on each system.

    • Valid Red Hat Subscription: Obtain Red Hat Enterprise Linux server subscriptions and apply one to each system.
    • CPUs: Two or more virtual CPUs
    • RAM: 4GB for each A and B system; 8GB for each C system
    • Disk space: About 20GB of disk space for each A and B system (10GB for the operating system and 10GB for docker storage). At least 30GB of disk space for C systems (or more depending on required container storage).

3.2. Set up Load Balancer and Database

On the first two systems (q01 and q02), install the haproxy load balancer and postgresql database. Haproxy will be configured as the access point and load balancer for the following services running on other systems:

  • Quay (ports 80 and 443 on B systems)
  • Redis (port 6379 on B systems)
  • RADOS (port 7480 on C systems)

Because the services on the two systems run as containers, you also need the docker service running. Here’s how to set up the A systems:

  1. Install and start docker service: Install, start, and enable the docker service.
  2. Open ports for haproxy service: Open all haproxy ports in SELinux and selected haproxy ports in the firewall:

    # setsebool -P haproxy_connect_any=on
    # firewall-cmd --permanent --zone=public --add-port=6379/tcp --add-port=7480/tcp
    success
    # firewall-cmd --reload
    success
  3. Set up haproxy service: Configure the /etc/haproxy/haproxy.cfg to point to the systems and ports providing the Quay, Redis, and Ceph RADOS services. Here are examples of defaults and added frontend and backend settings:

    #---------------------------------------------------------------------
    # common defaults that all the 'listen' and 'backend' sections will
    # use if not designated in their block
    #---------------------------------------------------------------------
    defaults
        mode                    tcp
        log                     global
        option                  httplog
        option                  dontlognull
        option http-server-close
        option forwardfor       except 127.0.0.0/8
        option                  redispatch
        retries                 3
        timeout http-request    10s
        timeout queue           1m
        timeout connect         10s
        timeout client          1m
        timeout server          1m
        timeout http-keep-alive 10s
        timeout check           10s
        maxconn                 3000
    
    #---------------------------------------------------------------------
    # main frontend which proxys to the backends
    #---------------------------------------------------------------------
    
    frontend  fe_http *:80
        default_backend             be_http
    frontend  fe_https *:443
        default_backend             be_https
    frontend fe_redis *:6379
       default_backend              be_redis
    frontend  fe_rdgw *:7480
        default_backend             be_rdgw
    backend be_http
        balance     roundrobin
        server quay01 quay01:80 check
        server quay02 quay02:80 check
        server quay03 quay03:80 check
    backend be_https
        balance     roundrobin
        server quay01 quay01:443 check
        server quay02 quay02:443 check
        server quay03 quay03:443 check
    backend be_rdgw
        balance     roundrobin
        server ceph01 ceph01:7480 check
        server ceph02 ceph02:7480 check
        server ceph03 ceph03:7480 check
    backend be_redis
    server quay01 quay01:6380 check inter 1s
    server quay02 quay02:6380 check inter 1s
    server quay03 quay03:6380 check inter 1s

    Once the new haproxy.cfg file is in place, restart the haproxy service.

    # systemctl restart haproxy
  4. Install / Deploy a Database: Install, enable and start the PostgreSQL database container. The following commands will:

    • Start the PostgreSQL database with the user, password and database all set. Data from the container will be stored on the host system in the /var/lib/pgsql/data directory.
    • List available extensions.
    • Create the pg_trgm extension.
    • Confirm the extension is installed

      $ mkdir -p /var/lib/pgsql/data
      $ chmod 777 /var/lib/pgsql/data
      $ sudo docker run -d --name postgresql_database \
          -v /var/lib/pgsql/data:/var/lib/pgsql/data:Z  \
          -e POSTGRESQL_USER=quayuser -e POSTGRESQL_PASSWORD=quaypass \
          -e POSTGRESQL_DATABASE=quaydb -p 5432:5432 \
          rhscl/postgresql-96-rhel7
      
      $ sudo docker exec -it postgresql_database /bin/bash -c 'echo "SELECT * FROM pg_available_extensions" | /opt/rh/rh-postgresql96/root/usr/bin/psql'
         name    | default_version | installed_version |           comment
      -----------+-----------------+-------------------+----------------------------------------
       adminpack | 1.0             |                   | administrative functions for PostgreSQL
      ...
      
      
      $ sudo docker exec -it postgresql_database /bin/bash -c 'echo "CREATE EXTENSION pg_trgm" | /opt/rh/rh-postgresql96/root/usr/bin/psql'
      CREATE EXTENSION
      
      $ sudo docker exec -it postgresql_database /bin/bash -c 'echo "SELECT * FROM pg_extension" | /opt/rh/rh-postgresql96/root/usr/bin/psql'
       extname | extowner | extnamespace | extrelocatable | extversion | extconfig | extcondition
      ---------+----------+--------------+----------------+------------+-----------+--------------
       plpgsql |       10 |           11 | f              | 1.0        |           |
       pg_trgm |       10 |         2200 | t              | 1.3        |           |
      (2 rows)
      
      $ sudo docker exec -it postgresql_database /bin/bash -c 'echo "ALTER USER quayuser WITH SUPERUSER;" | /opt/rh/rh-postgresql96/root/usr/bin/psql'
      ALTER ROLE
  5. Open the firewall: If you have a firewalld service active on your system, run the following commands to make the PostgreSQL port available through the firewall:

    # firewall-cmd --permanent --zone=trusted --add-port=5432/tcp
    success
    # firewall-cmd --reload
    success
  6. Test PostgreSQL Connectivity: Use the psql command to test connectivity to the PostgreSQL database. Try this on a remote system as well, to make sure you can access the service remotely:

    # yum install postgresql -y
    
    # psql -h localhost quaydb quayuser
    Password for user test:
    psql (9.2.23, server 9.6.5)
    WARNING: psql version 9.2, server version 9.6.
             Some psql features might not work.
    Type "help" for help.
    
    test=> \q

3.3. Set Up Ceph

For this Red Hat Quay configuration, we create a three-node Ceph cluster, with several other supporting nodes, as follows:

  • ceph01, ceph02, and ceph03 - Ceph Monitor, Ceph Manager and Ceph OSD nodes
  • ceph04 - Ceph RGW node
  • ceph05 - Ceph Ansible administration node

For details on installing Ceph nodes, see Installing Red Hat Ceph Storage on Red Hat Enterprise Linux.

Once you have set up the Ceph storage cluster, create a Ceph Object Gateway (also referred to as a RADOS gateway). See Installing the Ceph Object Gateway for details.

3.3.1. Install each Ceph node

On ceph01, ceph02, ceph03, ceph04, and ceph05, do the following:

  1. Review prerequisites for setting up Ceph nodes in Requirements for Installing Red Hat Ceph Storage. In particular:

  2. Prepare OSD storage (ceph01, ceph02, and ceph03 only). Set up the OSD storage on the three OSD nodes (ceph01, ceph02, and ceph03). See OSD Ansible Settings in Table 3.2 for details on supported storage types that you will enter into your Ansible configuration later. For this example, a single, unformatted block device (/dev/sdb), that is separate from the operating system, is configured on each of the OSD nodes. If you are installing on metal, you might want to add an extra hard drive to the machine for this purpose.
  3. Install Red Hat Enterprise Linux Server edition, as described in the RHEL 7 Installation Guide.
  4. Register and subscribe each Ceph node as described in the Registering Red Hat Ceph Storage Nodes. Here is how to subscribe to the necessary repos:

    # subscription-manager repos --disable=*
    # subscription-manager repos --enable=rhel-7-server-rpms
    # subscription-manager repos --enable=rhel-7-server-extras-rpms
    # subscription-manager repos --enable=rhel-7-server-rhceph-3-mon-rpms
    # subscription-manager repos --enable=rhel-7-server-rhceph-3-osd-rpms
    # subscription-manager repos --enable=rhel-7-server-rhceph-3-tools-rpms
  5. Create an ansible user with root privilege on each node. Choose any name you like. For example:

    # USER_NAME=ansibleadmin
    # useradd $USER_NAME -c "Ansible administrator"
    # passwd $USER_NAME
    New password: *********
    Retype new password: *********
    # cat << EOF >/etc/sudoers.d/admin
    admin ALL = (root) NOPASSWD:ALL
    EOF
    # chmod 0440 /etc/sudoers.d/$USER_NAME

3.3.2. Configure the Ceph Ansible node (ceph05)

Log into the Ceph Ansible node (ceph05) and configure it as follows. You will need the ceph01, ceph02, and ceph03 nodes to be running to complete these steps.

  1. In the Ansible user’s home directory create a directory to store temporary values created from the ceph-ansible playbook

    # USER_NAME=ansibleadmin
    # sudo su - $USER_NAME
    [ansibleadmin@ceph05 ~]$ mkdir ~/ceph-ansible-keys
  2. Enable password-less ssh for the ansible user. Run ssh-keygen on ceph05 (leave passphrase empty), then run and repeat ssh-copy-id to copy the public key to the Ansible user on ceph01, ceph02, and ceph03 systems:

    # USER_NAME=ansibleadmin
    # sudo su - $USER_NAME
    [ansibleadmin@ceph05 ~]$ ssh-keygen
    [ansibleadmin@ceph05 ~]$ ssh-copy-id $USER_NAME@ceph01
    [ansibleadmin@ceph05 ~]$ ssh-copy-id $USER_NAME@ceph02
    [ansibleadmin@ceph05 ~]$ ssh-copy-id $USER_NAME@ceph03
    [ansibleadmin@ceph05 ~]$ exit
    #
  3. Install the ceph-ansible package:

    # yum install ceph-ansible
  4. Create a symbolic between these two directories:

    # ln -s /usr/share/ceph-ansible/group_vars \
        /etc/ansible/group_vars
  5. Create copies of Ceph sample yml files to modify:

    # cd /usr/share/ceph-ansible
    # cp group_vars/all.yml.sample group_vars/all.yml
    # cp group_vars/osds.yml.sample group_vars/osds.yml
    # cp site.yml.sample site.yml
  6. Edit the copied group_vars/all.yml file. See General Ansible Settings in Table 3.1 for details. For example:

    ceph_origin: repository
    ceph_repository: rhcs
    ceph_repository_type: cdn
    ceph_rhcs_version: 3
    monitor_interface: eth0
    public_network: 192.168.122.0/24

    Note that your network device and address range may differ.

  7. Edit the copied group_vars/osds.yml file. See the OSD Ansible Settings in Table 3.2 for details. In this example, the second disk device (/dev/sdb) on each OSD node is used for both data and journal storage:

    osd_scenario: collocated
    devices:
      - /dev/sdb
    dmcrypt: true
    osd_auto_discovery: false
  8. Edit the /etc/ansible/hosts inventory file to identify the Ceph nodes as Ceph monitor, OSD and manager nodes. In this example, the storage devices are identified on each node as well:

    [mons]
    ceph01
    ceph02
    ceph03
    
    [osds]
    ceph01 devices="[ '/dev/sdb' ]"
    ceph02 devices="[ '/dev/sdb' ]"
    ceph03 devices="[ '/dev/sdb' ]"
    
    [mgrs]
    ceph01 devices="[ '/dev/sdb' ]"
    ceph02 devices="[ '/dev/sdb' ]"
    ceph03 devices="[ '/dev/sdb' ]"
  9. Add this line to the /etc/ansible/ansible.cfg file, to save the output from each Ansible playbook run into your Ansible user’s home directory:

    retry_files_save_path = ~/
  10. Check that Ansible can reach all the Ceph nodes you configured as your Ansible user:

    # USER_NAME=ansibleadmin
    # sudo su - $USER_NAME
    [ansibleadmin@ceph05 ~]$ ansible all -m ping
    ceph01 | SUCCESS => {
        "changed": false,
        "ping": "pong"
    }
    ceph02 | SUCCESS => {
        "changed": false,
        "ping": "pong"
    }
    ceph03 | SUCCESS => {
        "changed": false,
        "ping": "pong"
    }
    [ansibleadmin@ceph05 ~]$
  11. Run the ceph-ansible playbook (as your Ansible user):

    [ansibleadmin@ceph05 ~]$ cd /usr/share/ceph-ansible/
    [ansibleadmin@ceph05 ~]$ ansible-playbook site.yml

    At this point, the Ansible playbook will check your Ceph nodes and configure them for the services you requested. If anything fails, make needed corrections and rerun the command.

  12. Log into one of the three Ceph nodes (ceph01, ceph02, or ceph03) and check the health of the Ceph cluster:

    # ceph health
    HEALTH_OK
  13. On the same node, verify that monitoring is working using rados:

    # ceph osd pool create test 8
    # echo 'Hello World!' > hello-world.txt
    # rados --pool test put hello-world hello-world.txt
    # rados --pool test get hello-world fetch.txt
    # cat fetch.txt
    Hello World!

3.3.3. Install the Ceph Object Gateway

On the Ansible system (ceph05), configure a Ceph Object Gateway to your Ceph Storage cluster (which will ultimately run on ceph04). See Installing the Ceph Object Gateway for details.

3.4. Set up Quay and Redis

With Red Hat Enterprise Linux server installed on each of the three Quay systems (quay01, quay02, and quay03), install and start the Red Hat Quay and Redis services.

Important

When you go to configure Red Hat Quay, only do so on one of the three quay0* systems. Once that is done, the procedure will have you copy that configuration to the other two systems running the Quay service.

  1. Setup Docker: Install, enable, and start the docker service as shown here (see Getting Docker in RHEL 7 for details):
  2. Install / Deploy Redis: Run Redis as a container on each of the three quay0* systems:

    # mkdir -p /mnt/hostredis
    # chmod 777 /mnt/hostredis
    # docker run -d --restart=always -p 6379:6379 \
        -v /mnt/hostredis:/var/lib/redis/data:Z \
        registry.access.redhat.com/rhscl/redis-32-rhel7
  3. Check redis connectivity: You can use the telnet command to test connectivity to the redis service. Type MONITOR (to begin monitoring the service) and QUIT to exit:

    # yum install telnet -y
    # telnet 192.168.122.99 6379
    Trying 192.168.122.99...
    Connected to 192.168.122.99.
    Escape character is '^]'.
    MONITOR
    +OK
    +1525703165.754099 [0 172.17.0.1:43848] "PING"
    QUIT
    +OK
    Connection closed by foreign host.
  4. Add Quay authentication: Set up authentication to Quay.io, so you can pull the Quay container, as described in Accessing Red Hat Quay without a CoreOS login
  5. Install / Deploy Quay: Start and set up Red Hat Quay on quay01, then start that same service on quay02 and quay03 (using the shared configuration file).

    Note

    Add -e DEBUGLOG=true to the docker run command line for the quay container to enable debug level logging.

    On each of the three quay0* systems, run Red Hat Quay as a container, as follows:

    # mkdir -p /mnt/quay/config /mnt/quay/storage
    # firewall-cmd --permanent --zone=trusted --add-port=80/tcp
    # firewall-cmd --permanent --zone=trusted --add-port=443/tcp
    # firewall-cmd --reload
    
    # docker run --restart=always -p 443:443 -p 80:80 \
       --privileged=true \
       -v /mnt/quay/config:/conf/stack \
       -v /mnt/quay/storage:/datastorage \
       -d quay.io/coreos/quay:v2.9.5

    Wait for the Quay service to come up, then proceed to Completing the Guided Setup.

    Note

    The quay container startup can take several minutes. Type, docker ps to see the container id and docker logs -f <containerid> if you want to watch the progress. It’s getting near completion when you see the container open the /etc/hosts file. When attempting to access the Guided Setup you might receive a "502 Bad Gateway" nginx message. If you do, wait a while longer and try again.