path: root/README.mdwn

Baserock project public infrastructure
======================================

This repository contains the definitions for all of the Baserock Project's
infrastructure. This includes every service used by the project, except for
the mailing lists (hosted by [Pepperfish]) and the wiki (hosted by
[Branchable]).

Some of these systems are Baserock systems. Other are Ubuntu or Fedora based.
Eventually we want to move all of these to being Baserock systems.

The infrastructure is set up in a way that parallels the preferred Baserock
approach to deployment. All files necessary for (re)deploying the systems
should be contained in this Git repository, with the exception of certain
private tokens (which should be simple to inject at deploy time).

[Pepperfish]: http://listmaster.pepperfish.net/cgi-bin/mailman/listinfo
[Branchable]: http://www.branchable.com/


General notes
-------------

When instantiating a machine that will be public, remember to give shell
access everyone on the ops team. This can be done using a post-creation
customisation script that injects all of their SSH keys. The SSH public
keys of the Baserock Operations team are collected in
`baserock-ops-team.cloud-config.`.

Ensure SSH password login is disabled in all systems you deploy! See:
<https://testbit.eu/is-ssh-insecure/> for why. The Ansible playbook
`admin/sshd_config.yaml` can ensure that all systems have password login
disabled.


Administration
--------------

You can use [Ansible] to automate tasks on the baserock.org systems.

To run a playbook:

    ansible-playbook -i hosts $PLAYBOOK.yaml

To run an ad-hoc command (upgrading, for example):

    ansible-playbook -i hosts fedora -m command -a 'sudo yum update -y'
    ansible-playbook -i hosts ubuntu -m command -a 'sudo apt-get update -y'

[Ansible]: http://www.ansible.com


Deployment to OpenStack
-----------------------

The intention is that all of the systems defined here are deployed to an
OpenStack cloud. The instructions here harcode some details about the specific
tenancy at [DataCentred](http://www.datacentred.io) that the Baserock project
uses. It should be easy to adapt them for other OpenStack hosts, though.

### Credentials

The instructions below assume you have the following environment variables set
according to the OpenStack host you are deploying to:

 - `OS_AUTH_URL`
 - `OS_TENANT_NAME`
 - `OS_USERNAME`
 - `OS_PASSWORD`

When using `morph deploy` to deploy to OpenStack, you will need to set these
variables, because currently Morph does not honour the standard ones. See:
<https://storyboard.baserock.org/#!/story/35>.

 - `OPENSTACK_USER=$OS_USERNAME`
 - `OPENSTACK_PASSWORD=$OS_PASSWORD`
 - `OPENSTACK_TENANT=$OS_TENANT_NAME`

The `location` field in the deployment .morph file will also need to point to
the correct `$OS_AUTH_URL`.

### Firewall / Security Groups

The instructions assume the presence of a set of security groups. You can
create these by running the following Ansible playbook. You'll need the
OpenStack Ansible modules cloned from
`https://github.com/openstack-ansible/openstack-ansible-modules/`.

    ANSIBLE_LIBRARY=../openstack-ansible-modules ansible-playbook -i hosts \
        firewall.yaml

### Placeholders

The commands below use a couple of placeholders like $network_id, you can set
them in your environment to allow you to copy and paste the commands below
as-is.

 - `export fedora_image_id=...` (find this with `glance image-list`)
 - `export network_id=...` (find this with `neutron net-list`)
 - `export keyname=...` (find this with `nova keypair-list`)

The `$fedora_image_id` should reference a Fedora Cloud image. You can import
these from <http://www.fedoraproject.org/>. At time of writing, these
instructions were tested with Fedora Cloud 21 for x86_64.

Backups
-------

Backups of git.baserock.org's data volume are run by and stored on on a
Codethink-managed machine named 'access'. They will need to migrate off this
system before long.  The backups are taken without pausing services or
snapshotting the data, so they will not be 100% clean. The current
git.baserock.org data volume does not use LVM and cannot be easily snapshotted.

Backups of 'gerrit' and 'database' are handled by the
'baserock_backup/backup.py' script. This currently runs on an instance in
Codethink's internal OpenStack cloud.

Instances themselves are not backed up. In the event of a crisis we will
redeploy them from the infrastructure.git repository. There should be nothing
valuable stored outside of the data volumes that are backed up.

To prepare the infrastructure to run the backup scripts you will need to run
the following playbooks:

    ansible-playbook -i hosts baserock_frontend/instance-backup-config.yml
    ansible-playbook -i hosts baserock_database/instance-backup-config.yml
    ansible-playbook -i hosts baserock_gerrit/instance-backup-config.yml

NOTE: to run these playbooks you need to have the public ssh key of the backups
instance in `keys/backup.key.pub`.


Deployment with Packer
----------------------

> **NOTE**: I no longer think that Packer is the right tool for our needs. This
> is partly because of critical bugs that have not been fixed since I started
> using it (e.g. <https://github.com/mitchellh/packer/issues/1462>), and partly
> because I realised that I was just using it to wrap `nova` and
> `ansible-playbook`, and it is simple enough to use those commands directly.
>
> I had hoped that we could make use of Packer's multiple backends in order to
> test systems locally in Docker before deploying them to OpenStack. It turns
> out Docker is sufficiently different to OpenStack that this doesn't make life
> any easier during development. Networking setup is different, systemd doesn't
> work inside Docker by default, base images are different in other ways, etc.
>
> So I recommend not using Packer for future systems, and I will try to
> migrate the definitions for the existing ones to just use Ansible.
>
> Sam Thursfield 10/04/15

Some of the systems are built with [Packer]. I chose Packer because it provides
similar functionality to the `morph deploy` command, although its
implementation makes different tradeoffs. The documentation below shows the
commands you need to run to build systems with Packer. Some of the systems can
be deployed as Docker images as well as OpenStack images, to enable local
development and testing.

The following error from Packer means that you didn't set your credentials
correctly in the `OS_...` environment variables, or they were not accepted.

> Build 'production' errored: Missing or incorrect provider

The the Packer tool requires a floating IP to be available at the time a system
is being deployed to OpenStack. Currently 185.43.218.169 should be used for
this.  If you specify a floating IP that is in use by an existing instance, you
will steal it for your own instance and probably break one of our web services.

[Packer]: http://www.packer.io/


Systems
-------

### Front-end

The front-end provides a reverse proxy, to allow more flexible routing than
simply pointing each subdomain to a different instance using separate public
IPs. It also provides a starting point for future load-balancing and failover
configuration.

To deploy this system:

    nova boot frontend-haproxy \
        --key-name=$keyname \
        --flavor=dc1.1x1.1 \
        --image=$fedora_image_id \
        --nic="net-id=$network_id" \
        --security-groups default,gerrit,web-server \
        --user-data ./baserock-ops-team.cloud-config
    ansible-playbook -i hosts baserock_frontend/image-config.yml
    ansible-playbook -i hosts baserock_frontend/instance-config.yml

    ansible -i hosts -m service -a 'name=haproxy enabled=true state=started' \
        --sudo frontend-haproxy

The baserock_frontend system is stateless.

Full HAProxy 1.5 documentation: <https://cbonte.github.io/haproxy-dconv/configuration-1.5.html>.

If you want to add a new service to the Baserock Project infrastructure via
the frontend, do the following:

- request a subdomain that points at 185.43.218.170 (frontend)
- alter the haproxy.cfg file in the baserock_frontend/ directory in this repo
  as necessary to proxy requests to the real instance
- run the baserock_frontend/instance-config.yml playbook
- run `ansible -i hosts -m service -a 'name=haproxy enabled=true
  state=restarted' --sudo frontend-haproxy`

OpenStack doesn't provide any kind of internal DNS service, so you must put the
fixed IP of each instance.

### Database

Baserock infrastructure uses a shared [MariaDB] database. MariaDB was chosen
because Storyboard only supports MariaDB.

To deploy this system to production:

    nova boot database-mariadb \
        --key-name=$keyname \
        --flavor dc1.1x1 \
        --image=$fedora_image_id \
        --nic="net-id=$network_id,v4-fixed-ip=192.168.222.30" \
        --security-groups default,database-mysql \
        --user-data ./baserock-ops-team.cloud-config
    nova volume-create \
        --display-name database-volume \
        --display-description 'Database volume' \
        --volume-type Ceph \
        100
    nova volume-attach database-mariadb <volume ID> /dev/vdb

    ansible-playbook -i hosts baserock_database/image-config.yml
    ansible-playbook -i hosts baserock_database/instance-config.yml

At this point, if you are restoring from a backup, rsync the data across
from your backup server on the instance, then start the mariadb service and you
are done.

    sudo --preserve-env -- rsync --archive --chown mysql:mysql --hard-links \
         --info=progress2 --partial --sparse \
         root@backupserver:/srv/backup/database/* /var/lib/mysql
    sudo systemctl enable mariadb.service
    sudo systemctl start mariadb.service

NOTE: If you see the following message in the journal:

    The datadir located at /var/lib/mysql needs to be upgraded using 'mysql_upgrade' tool. This can be done using the following steps

This is because the backup you are importing is from an older version of
MariaDB. To fix this, as the message says, you only need to run:

    sudo -u mysql mysql_upgrade -u root -p

If you are starting from scratch, you need to prepare the system by adding
the required users and databases. Run the following playbook, which can
be altered and rerun whenever you need to add more users or databases, or
you want to check the database configuration matches what you expect.

    ansible -i hosts -m service -a 'name=mariadb enabled=true state=started'
    ansible-playbook -i hosts baserock_database/instance-mariadb-config.yml

[MariaDB]: https://www.mariadb.org

### Mail relay

The mail relay is currently a Fedora Cloud 21 image running Exim. You should be
able to take a Fedora Cloud 21 base image, instantiate it in the
'internal-mail-relay' security group, and then run
'baserock_mail/instance-config.yml' to configure it and start the service.

It is configured to only listen on its internal IP. It's not intended to
receive mail, or relay mail sent by systems outside the baserock.org cloud.

### OpenID provider

To deploy a development instance:

    packer build -only=development baserock_openid_provider/packer_template.json
    baserock_openid_provider/develop.sh
    # Now you have a root shell inside your container
    cd /srv/baserock_openid_provider
    python ./manage.py runserver 0.0.0.0:80
    # Now you can browse to http://localhost:80/ and see the server.

To deploy this system to production:

    vim baserock_openid_provider/baserock_openid_provider/settings.py

Edit the DATABASES['default']['HOST'] to point to the fixed IP of the
'database' machine, and check the settings. See:
https://docs.djangoproject.com/en/1.7/howto/deployment/checklist/

    packer build -only=production baserock_openid_provider/packer_template.json

    nova boot openid.baserock.org \
        --key-name $keyname \
        --flavor dc1.1x1 \
        --image 'baserock_openid_provider' \
        --nic "net-id=$network_id',v4-fixed-ip=192.168.222.67" \
        --security-groups default,web-server
        --user-data ./baserock-ops-team.cloud-config

    ansible-playbook -i hosts baserock_openid_provider/instance-config.yml

To change Cherokee configuration, it's usually easiest to use the
cherokee-admin tool in a running instance. SSH in as normal but forward port
9090 to localhost (pass `-L9090:localhost:9090` to SSH). Backup the old
/etc/cherokee/cherokee.conf file, then run `cherokee-admin`, and log in using
the creditials it gives you. After changing the configuration, please update
the cherokee.conf in infrastructure.git to match the changes `cherokee-admin`
made.

### Gerrit

To deploy to production, run these commands in a Baserock 'devel'
or 'build' system.

    nova volume-create \
        --display-name gerrit-volume \
        --display-description 'Gerrit volume' \
        --volume-type Ceph \
        100

    git clone git://git.baserock.org/baserock/baserock/infrastructure.git
    cd infrastructure

    morph build systems/gerrit-system-x86_64.morph
    morph deploy baserock_gerrit/baserock_gerrit.morph

    nova boot gerrit.baserock.org \
        --key-name $keyname \
        --flavor 'dc1.2x4.40' \
        --image baserock_gerrit \
        --nic "net-id=$network_id,v4-fixed-ip=192.168.222.69" \
        --security-groups default,gerrit,git-server,web-server \
        --user-data baserock-ops-team.cloud-config

    nova volume-attach gerrit.baserock.org <volume-id> /dev/vdb

Accept the license and download the latest Java Runtime Environment from
http://www.oracle.com/technetwork/java/javase/downloads/server-jre8-downloads-2133154.html

Accept the license and download the latest Java Cryptography Extensions from
http://www.oracle.com/technetwork/java/javase/downloads/jce8-download-2133166.html

Save these two files in the baserock_gerrit/ folder. The instance-config.yml
Ansible playbook will upload them to the new system.

    # Don't copy-paste this! Use the Oracle website instead!
    wget --no-cookies --no-check-certificate \
         --header "Cookie: gpw_e24=http%3A%2F%2Fwww.oracle.com%2F; oraclelicense=accept-securebackup-cookie" \
         "http://download.oracle.com/otn-pub/java/jdk/8u40-b25/server-jre-8u40-linux-x64.tar.gz" 
    wget --no-cookies --no-check-certificate \
          --header "Cookie: gpw_e24=http%3A%2F%2Fwww.oracle.com%2F; oraclelicense=accept-securebackup-cookie" \
          "http://download.oracle.com/otn-pub/java/jce/8/jce_policy-8.zip"

    ansible-playbook -i hosts baserock_gerrit/instance-config.yml

For baserock.org Gerrit you will also need to run:

    ansible-playbook -i hosts baserock_gerrit/instance-ca-certificate-config.yml

If you are restoring from a backup, rsync the data across from your
backup server on the instance, then start the gerrit service.

    systemctl stop gerrit.service
    rm -r /srv/gerrit/*

    rsync --archive --chown gerrit:gerrit --hard-links \
         --info=progress2 --partial --sparse \
         root@backupserver:/srv/backup/gerrit/* /srv/gerrit/

    systemctl start gerrit.service

NOTE: If you are restoring a backup from an older version of Gerrit, you
might need to run some of the following commands to migrate the schemas of the
database, and also gerrit data (This was needed to move from 2.9.4 to 2.11.4):

    java -jar /opt/gerrit/gerrit-2.11.3.war init -d /srv/gerrit
    java -jar /opt/gerrit/gerrit-2.11.3.war reindex -d /srv/gerrit


#### Access control

Gerrit should now be up and running and accessible through the web interface.
By default this is on port 8080. Log into the new Gerrit instance with your
credentials. Make sure you're the first one to have registered, and you will
automatically have been added to the Administrators group.

You can add more users into the Administrators group later on using the [gerrit
set-members] command, or the web interface.

Go to the settings page, 'HTTP Password' and generate a HTTP password for
yourself. You'll need it in the next step. The password can take a long time to
appear for some reason, or it might not work at all. Click off the page and
come back to it and it might suddenly have appeared. I've not investigated why
this happens.

Generate the SSH keys you need, if you don't have them.

    mkdir -p keys
    ssh-keygen -t rsa -b 4096 -C 'lorry@gerrit.baserock.org' -N '' -f keys/lorry-gerrit.key

Now set up the Gerrit access configuration. This Ansible playbook requires a
couple of non-standard packages.

    git clone git://git.baserock.org/delta/python-packages/pygerrit.git
    git clone git://github.com/ssssam/ansible-gerrit
    cd ansible-gerrit && make; cd -

    export GERRIT_URL=gerrit web URL
    export GERRIT_ADMIN_USERNAME=your username
    export GERRIT_ADMIN_PASSWORD=your generated HTTP password
    export GERRIT_ADMIN_REPO=ssh://you@gerrit:29418/All-Projects.git

    ANSIBLE_LIBRARY=./ansible-gerrit PYTHONPATH=./pygerrit \
    ansible-playbook baserock_gerrit/gerrit-access-config.yml

[gerrit set-members]: https://gerrit-documentation.storage.googleapis.com/Documentation/2.9.4/cmd-set-members.html

#### Mirroring

Run:

    ansible-playbook -i hosts baserock_gerrit/instance-mirroring-config.yml

Now clone the Gerrit's lorry-controller configuration repository, commit the
configuration file to it, and push.

    # FIXME: we could use the git_commit_and_push Ansible module for this now,
    # instead of doing it manually.

    git clone ssh://$GERRIT_ADMIN_USERNAME@gerrit.baserock.org:29418/local-config/lorries.git /tmp/lorries
    cp baserock_gerrit/lorry-controller.conf /tmp/lorries
    cd /tmp/lorries
    git checkout -b master
    git add .
    git commit -m "Add initial Lorry Controller mirroring configuration"
    git push origin master
    cd -

Now SSH in as 'root' to gerrit.baserock.org, tunnelling the lorry-controller
webapp's port to your local machine:

    ssh -L 12765:localhost:12765 root@gerrit.baserock.org

Visit <http://localhost/1.0/status-html>. You should see the lorry-controller
status page. Click 'Re-read configuration', if there are any errors in the
configuration it'll tell you. If not, it should start mirroring stuff from
your Trove.

Create a Gitano account on the Trove you want to push changes to for the Gerrit
user. The `instance-config.yml` Ansible playbook will have generated an SSH
key. Run these commands on the Gerrit instance:

    ssh git@git.baserock.org user add gerrit "gerrit.baserock.org" gerrit@baserock.org
    ssh git@git.baserock.org as gerrit sshkey add main < ~gerrit/.ssh/id_rsa.pub

Add the 'gerrit' user to the necessary -writers groups on the Trove, to allow
the gerrit-replication plugin to push merged changes to 'master' in the Trove.

    ssh git@git.baserock.org group adduser baserock-writers gerrit
    ssh git@git.baserock.org group adduser local-config-writers gerrit

Add the host key of the remote trove, to the Gerrit system:

    sudo -u gerrit sh -c 'ssh-keyscan git.baserock.org >> ~gerrit/.ssh/known_hosts'

Check the 'gerrit' user's Trove account is working.

    sudo -u gerrit ssh git@git.baserock.org whoami

Now enable the gerrit-replication plugin, check that it's now in the list of
plugins, and manually start a replication cycle. You should see log output from
the final SSH command showing any errors.

    ssh $GERRIT_ADMIN_USERNAME@gerrit.baserock.org -p 29418 gerrit plugin enable replication
    ssh $GERRIT_ADMIN_USERNAME@gerrit.baserock.org -p 29418 gerrit plugin ls
    ssh $GERRIT_ADMIN_USERNAME@gerrit.baserock.org -p 29418 replication start --all --wait

### Storyboard

We use a slightly adapted version of
<https://github.com/openstack-infra/puppet-storyboard> to deploy Storyboard.

There's no development deployment for Storyboard at this time: the Puppet
script expects to start services using systemd, and that doesn't work by
default in a Docker container.

To deploy the production version:

    packer build -only=production baserock_storyboard/packer_template.json
    nova boot openid_provider
        --flavor dc1.1x1 --image 'baserock_storyboard' \
        --key-name=$keyname storyboard.baserock.org \
        --nic="net-id=$network_id"
        --security-groups default,web-server
        --user-data baserock-ops-team.cloud-config

Storyboard deployment does not yet work fully (you can manually kludge it into
working after deploying it, though).

### Masons

Mason is the name we use for an automated build and test system used in the
Baserock project. The V2 Mason that runs as <https://mason-x86-32.baserock.org/>
and <https://mason-x86-64.baserock.org/> lives in definitions.git, and is thus
available in infrastructure.git too by default.

To build mason-x86-64:

    git clone git://git.baserock.org/baserock/baserock/infrastructure.git
    cd infrastructure

    morph build systems/build-system-x86_64.morph
    morph deploy baserock_mason_x86_64/mason-x86-64.morph

    nova boot mason-x86-64.baserock.org \
        --key-name $keyname \
        --flavor 'dc1.2x2' \
        --image baserock_mason_x86_64 \
        --nic "net-id=$network_id,v4-fixed-ip=192.168.222.80" \
        --security-groups internal-only,mason-x86
        --user-data baserock-ops-team.cloud-config

The mason-x86-32 system is the same, just subsitute '64' for '32' in the above
commands.

Note that the Masons are NOT in the 'default' security group, they are in
'internal-only'. This is a way of enforcing the [policy] that the Baserock
reference system definitions can only use source code hosted on
git.baserock.org, by making it impossible to fetch code from anywhere else.

[policy]: http://wiki.baserock.org/policies/