Synnefo

README.deploy -- Instructions for a basic Synnefo deployment

This document describes the basic steps to obtain a basic, working Synnefo

deployment. It begins by examining the different node roles, then moves to the

installation and setup of distinct software components.

It is current as of Synnefo v0.5.2.

Node types

===========

Nodes in a Synnefo deployment belong in one of the following types:

 * DB:

   A node [or more than one nodes, if using an HA configuration], running a DB

   engine supported by the Django ORM layer. The DB is the single source of

   truth for the servicing of API requests by Synnefo.

   Services: PostgreSQL / MySQL

 * APISERVER:

   A node running the implementation of the OpenStack API, in Django. Any number

   of APISERVERs can be used, in a load-balancing configuration, without any

   special consideration. Access to a common DB ensures consistency.

   Services: Web server, vncauthproxy

 * QUEUE:

   A node running the RabbitMQ software, which provides AMQP functionality. More

   than one QUEUE nodes may be deployed, in an HA configuration. Such

   deployments require shared storage, provided e.g., by DRBD.

   Services: RabbitMQ [rabbitmq-server]

 * LOGIC:

   A node running the business logic of Synnefo, in Django. It dequeues

   messages from QUEUE nodes, and provides the context in which business logic

   functions run. It uses Django ORM to connect to the common DB and update the

   state of the system, based on notifications received from the rest of the

   infrastructure, over AMQP.

   Services: the Synnefo logic dispatcher [/logic/dispatcher.py]

 * GANETI-MASTER and GANETI-NODE:

   A single GANETI-MASTER and a large number of GANETI-NODEs constitute the

   Ganeti backend for Synnefo, which undertakes all VM management functions.

   Any APISERVER can issue commands to the GANETI-MASTER, over RAPI, to effect

   changes in the state of the VMs. The GANETI-MASTER runs the Ganeti request

   queue.

   Services:

     only on GANETI-MASTER:

       the Synnefo Ganeti monitoring daemon [/ganeti/snf-ganeti-eventd]

       the Synnefo Ganeti hook [/ganeti/snf-ganeti-hook.py].

     on each GANETI_NODE:

       a deployment-specific KVM ifup script

       properly configured NFDHCPD

Installation Process

=====================

This section describes the installation process of the various node roles in a

Synnefo deployment.

0. Allocation of physical nodes:

   Determine the role of every physical node in your deployment.

1. Ganeti installation:

   Synnefo requires a working Ganeti installation at the backend. Installation

   of Ganeti is not covered by this document, please refer to

   http://docs.ganeti.org/ganeti/current/html for all the gory details. A

   successful Ganeti installation concludes with a working GANETI-MASTER and a

   number of GANETI-NODEs.

2. RabbitMQ installation:

   RabbitMQ is used as a generic message broker for the system. It should be

   installed on two seperate QUEUE nodes (VMs should be enough for the moment)

   in a high availability configuration as described here:

     http://www.rabbitmq.com/pacemaker.html

   After installation, create a user and set its permissions

     rabbitmqctl add_user okeanos 0k3@n0s

     rabbitmqctl set_permissions -p / okeanos  "^.*" ".*" ".*"

   The values set for the user and password must be mirrored in the

   RABBIT_* variables in settings.py (see step 6)

3. Web server installation:

   A Web Server (e.g., Apache) needs to be installed on the APISERVERs,

   and be configured to run the Synnefo Django project appropriately. Selection

   and configuration of a Web server is outside the scope of this document.

   For testing or development purposes, Django's own development server,

   `./manage.py runserver' can be used.

4. Installation of the Synnefo Django project:

   As of v0.5 the Synnefo Django project needs to be installed on nodes

   of type APISERVER, and LOGIC, with a properly configured settings.py. In

   later revisions, the specific parts of the Django project which need to run

   on each node type will be identified.

   Synnefo is written in Python 2.6 and depends on the following Python modules:

   [package versions confirmed to be compatible are in braces]

    * django 1.2 [Django==1.2.4]

    * simplejson [simplejson==2.1.3]

    * pycurl [pycurl==7.19.0]

    * python-dateutil  [python-dateutil==1.4.1]

      WARNING: version python-dateutil==2.0 downloaded by pip known *not* to

               work with Python 2.6

    * python-ipy [IPy==0.75]

        also verified to work with python-ipy 0.70-1 as shipped with Squeeze

    * south [south==0.7.1]

      WARNING: might not work with Debian Squeeze's default south-0.7-1 package.

    * amqplib [amqplib==0.6.1]

    * lockfile [lockfile==0.8]

    * python-daemon [python-daemon==1.5.5]

    * python-prctl [python-prctl==1.3.0]

   also, depending on the database engine of choice, on one of the following:

    * MySQL-python [MySQL-python==1.2.3]

    * psycopg2 [psycopg2==2.4]

   if the invitations application is deployed, the following dependencies should

   be installed:

    * pycrypto==2.1.0

   To run the user interface tests, selenium must be installed

    * selenium [?]

   The easiest method for installation of the Django project is to setup a

   working environment through virtualenv. Alternatively, you can use your

   system's package manager to install the dependencies (e.g. Macports has them

   all).

   * On Snow Leopard and linux (64-bit), you have to set the following

     environment variable for pip to compile the dependencies correctly.

  	   $ export ARCHFLAGS="-arch x86_64"

   * On Ubuntu, a few more packages must be installed before installing the

     prerequisite Python libraries

	   $ sudo aptitude install libcurl3-gnutls libcurl3-gnutls-dev uuid-dev

   Checkout the code and install the Python prerequisites. This assumes that

   python is already installed on the host.

    $ sudo easy_install virtualenv

    $ git clone https://user@code.grnet.gr/git/synnefo synnefo

    $ virtualenv --python=python2.6 synnefo --no-site-packages

    ...

    $ cd synnefo

    $ ./bin/pip install <list_of_dependencies>

    [WARNING]: The software must be checked out in a directory named synnefo,

    otherwise python imports will not work. Therefore, do not change the

    or rename the checkout path.

5. Database installation:

   A database supported by the Django ORM layer must be installed on nodes

   of type DB. The choices are: SQLIte, MySQL, PostgreSQL.

   * SQLite:

     The python sqlite driver is available by default with Python so no

     additional configuration is required. Also, most self-respecting systems

     have the sqlite library installed by default.

   * MySQL:

      MySQL must be installed first:

      * Ubuntu - Debian

	      $ sudo apt-get install libmysqlclient-dev

      * MacPorts

	      $ sudo port install mysql5

      Install the MySQL python library on servers running the Django project:

	    $ bin/pip install MySQL-python

      Note: On MacOSX with Mysql install from MacPorts the above command will

            fail complaining that it cannot find the mysql_config command. Do

            the following and restart the installation

	        $ echo "mysql_config = /opt/local/bin/mysql_config5" >> \

                                         ./build/MySQL-python/site.cfg

      Configure a MySQL db/account for synnefo

	    $ mysql -u root -p

    	mysql> create database synnefo;

	    mysql> show databases;

	    mysql> GRANT ALL on synnefo.* TO username IDENTIFIED BY 'password';

     IMPORTANT:

        MySQL *must* be set in READ-COMMITED mode, e.g. by setting

        transaction-isolation = READ-COMMITTED

        in the [mysqld] section of /etc/mysql/my.cnf.

        Alternatively, make sure the following code fragment stays enabled

        in settings.d/10-database.conf:

            if DATABASES['default']['ENGINE'].endswith('mysql'):

                DATABASES['default']['OPTIONS'] = {

                        'init_command': 'SET storage_engine=INNODB; ' +

                            'SET SESSION TRANSACTION ISOLATION LEVEL READ COMMITTED',

                }

   * PostgreSQL

     You need to install the PostgreSQL binaries:

     * Ubuntu - Debian

	     $ sudo apt-get install postgresql-8.4 libpq-dev

     * MacPorts

	     $ sudo port install postgresql84

     Install the postgres Python library

	    $ bin/pip install psycopg2

     Configure a postgres db/account for synnefo:

     Become the postgres user, connect to PostgreSQL:

       $ sudo su - postgres

       $ psql

	 Run the following commands:

	   DROP DATABASE synnefo;

	   DROP USER username;

	   CREATE USER username WITH PASSWORD 'password';

	   CREATE DATABASE synnefo;

	   GRANT ALL PRIVILEGES ON DATABASE synnefo TO username;

	   ALTER DATABASE synnefo OWNER TO username;

	   ALTER USER username CREATEDB;

     The last line enables the newly created user to create own databases. This

     is needed for Django to create and drop the test_synnefo database for unit

     testing.

6. Setting up the Django project:

   The settings.py file for Django may be derived by concatenating the

   settings.py.dist file contained in the Synnefo distribution with a file

   containing custom modifications, which shall override all settings deviating

   from the supplied settings.py.dist. This is recommended to minimize the load

   of reconstructing settings.py from scratch, since each release currently

   brings heavy changes to settings.py.dist.

   Add the following to your custom settings.py, depending on your choice

   of DB:

   * SQLite

	 PROJECT_PATH = os.path.dirname(os.path.abspath(__file__)) + '/'

	 DATABASES = {

	     'default': {

		     'ENGINE': 'django.db.backends.sqlite3',

		     'NAME': PROJECT_PATH + 'synnefo.db' # WARN: This must be an absolute path

	     }

	 }

   * MySQL

 	 DATABASES = {

	     'default': {

             'ENGINE': 'django.db.backends.mysql',

             'NAME': 'synnefo',

             'USER': 'USERNAME',

             'PASSWORD': 'PASSWORD',

             'HOST': 'HOST',

             'PORT': 'PORT',

             'OPTIONS': {

                 'init_command': 'SET storage_engine=INNODB',

             }

	    }

	}

   * PostgreSQL

     DATABASES = {

	     'default': {

             'ENGINE': 'django.db.backends.postgresql_psycopg2',

             'NAME': 'DATABASE',

             'USER': 'USERNAME',

             'PASSWORD': 'PASSWORD',

             'HOST': 'HOST',

             'PORT': 'PORT',

	     }

     }

    Try it out. The following command will attempt to connect to the DB and

    print out DDL statements. It should not fail.

	$ ./bin/python manage.py sql db

7. Initialization of Synnefo DB:

   You need to initialize the Synnefo DB and load fixtures

   db/fixtures/{flavors,images}.json, which make the API usable by end users by

   defining a sample set of hardware configurations (flavors) and OS images.

   IMPORTANT: Be sure to modify db/fixtures/initial_data.json and select

   a unique token for each of the initial and any other users defined in this

   file. DO NOT LEAVE THE SAMPLE AUTHENTICATION TOKENS enabled in deployed

   configurations.

     $ ./bin/python manage.py syncdb

     $ ./bin/python manage.py migrate db

     $ ./bin/python manage.py loaddata db/fixtures/flavors.json

     $ ./bin/python manage.py loaddata db/fixtures/images.json

8. Finalization of settings.py:

   Set the BACKEND_PREFIX_ID variable to some unique prefix, e.g. your commit

   username in settings.py. Several functional conventions within the system

   require this variable to include a dash at its end (e.g. snf-)

9. Installation of the Ganeti monitoring daemon, /ganeti/snf-ganeti-eventd:

   The Ganeti monitoring daemon must run on GANETI-MASTER.

   The monitoring daemon is configured through /etc/synnefo/settings.conf.

   An example is provided under snf-ganeti-tools/.

   If run from the repository directory, make sure to have snf-ganeti-tools/

   in the PYTHONPATH.

   You may also build Debian packages directly from the repository:

   $ cd snf-ganeti-tools

   $ dpkg-buildpackage -b -uc -us

   # dpkg -i ../snf-ganeti-tools-*deb

   TBD: how to handle master migration.

10. Installation of the Synnefo dispatcher, /logic/dispatcher.py:

    The logic dispatcher is part of the Synnefo Django project and must run

    on LOGIC nodes.

    The dispatcher retrieves messages from the queue and calls the appropriate

    handler function as defined in the queue configuration in `setttings.py'.

    The default configuration should work directly without any modifications.

    For the time being The dispatcher must be run by hand:

      $ ./bin/python ./logic/dispatcher.py

    The dispatcher should run in at least 2 instances to ensure high

    (actually, increased) availability.

11. Installation of the Synnefo Ganeti hook:

    The generic Synnefo Ganeti hook wrapper resides in the snf-ganeti-tools/

    directory of the Synnefo repository.

    The hook needs to be enabled for phases post-{add,modify,reboot,start,stop}

    by *symlinking* in

    /etc/ganeti/hooks/instance-{add,modify,reboot,start,stop}-post.d on

    GANETI-MASTER, e.g.:

    root@ganeti-master:/etc/ganeti/hooks/instance-start-post.d# ls -l

    lrwxrwxrwx 1 root root 45 May   3 13:45 00-snf-ganeti-hook -> /home/devel/synnefo/snf-ganeti-hook/snf-ganeti-hook.py

    IMPORTANT: The link name may only contain "upper and lower case, digits,

    underscores and hyphens. In other words, the regexp ^[a-zA-Z0-9_-]+$."

    See:

    http://docs.ganeti.org/ganeti/master/html/hooks.html?highlight=hooks#naming

    If run from the repository directory, make sure to have snf-ganeti-tools/

    in the PYTHONPATH.

    Alternative, build Debian packages which take care of building, installing

    and activating the Ganeti hook automatically, see step. 9.

12. Installation of the VNC authentication proxy, vncauthproxy:

    To support OOB console access to the VMs over VNC, the vncauthproxy

    daemon must be running on every node of type APISERVER.

    Download and install vncauthproxy from its own repository,

    at https://code.grnet.gr/git/vncauthproxy (known good commit: tag v1.0).

    Download and install a specific repository commit:

    $ bin/pip install -e git+https://code.grnet.gr/git/vncauthproxy@INSERT_COMMIT_HERE#egg=vncauthproxy

    Create /var/log/vncauthproxy and set its permissions appropriately.

    Alternatively, you can build Debian packages. To do so,

    checkout the "debian" branch of the vncauthproxy repository

    (known good commit: tag debian/v1.0):

    $ git checkout debian

    Then build debian package, and install as root:

    $ dpkg-buildpackage -b -uc -us

    # dpkg -i ../vncauthproxy_1.0-1_all.deb

    --Failure to build the package on the Mac.

    libevent, a requirement for gevent which in turn is a requirement for

    vncauthproxy is not included in MacOSX by default and installing it with

    MacPorts does not lead to a version that can be found by the gevent

    build process. A quick workaround is to execute the following commands:

    cd $SYNNEFO

    sudo pip install -e git+https://code.grnet.gr/git/vncauthproxy@5a196d8481e171a#egg=vncauthproxy

    <the above fails>

    cd build/gevent

    sudo python setup.py -I/opt/local/include -L/opt/local/lib build

    cd $SYNNEFO

    sudo pip install -e git+https://code.grnet.gr/git/vncauthproxy@5a196d8481e171a#egg=vncauthproxy

13. Installation of the customized Ganeti Instance Image for image deployment:

    For Synnefo to be able to launch VMs from specified Images, you need

    the gnt-instance-image OS Provider installed on *all* Ganeti nodes.

    There are 2 different ways to install gnt-instance-image on a node:

    1.As a debian package (recommended)

        Download and install the debian package found here:

        https://code.grnet.gr/attachments/download/351/ganeti-instance-image_0.5.1-1-snf1_all.deb

        You can do this by running the following:

        $ wget https://code.grnet.gr/attachments/download/351/ganeti-instance-image_0.5.1-1-snf1_all.deb

        # dpkg -i ganeti-instance-image_0.5.1-1-snf1_all.deb

        Make any additional configuration changes in 

        /etc/default/ganeti-instance-image.

        It is recommended to change the default IMAGE_DIR from 

        /var/cache/ganeti-instance-image to something like /srv/example_images_repo

	Be sure to have all the package's needed dependencies installed 

        on your system.

    2.From source    

        Download and install gnt-instance-image from its own repository,

        at https://code.grnet.gr/git/gnt-instance-image. It's

        recommended to use the win-support branch (known good commit:

        a7fb0e606623e7ac10d06ac935b0afe7f03d496c).

        Make sure to enable progress monitoring, using the --with-progress-monitor

        argument to configure. This requires the snf-progress-monitor tool,

        provided in snf-ganeti-tools/ and also as part of the snf-ganeti-tools

        Debian package.

        After installing gnt-instance-image do the following:

        1. $ cd /path-to-repo

           # cp ./defaults /etc/default/ganeti-instance-image

        2. In /etc/ganeti/instance-image/hooks, make sure the hooks you want to

           run during the instance creation process have execute permission.

           For linux you will need at lease `grub' and `root_passwd' to make the

           instance usable:

           chmod +x /etc/ganeti/instance-image/hooks/linux/{grub,root_passwd}

           For security reasons make sure `ssh' hook is also enabled.

           For windows you will need `mbr' and `admin_passwd':

           chmod +x /etc/ganeti/instance-image/hooks/windows/{mbr,admin_passwd}

           For both architectures it is also highly recommended to enable the

           `hostname' hook too:

           chmod +x /et/ganeti/instance-image/hooks/{linux,windows}/hostname

    Your custom Images should be stored in a dump format under

    /var/cache/ganeti-instance-image (default) or a different directory of your

    choice, accordingly set in /etc/default/ganeti-instance-image. The latter

    is recommended. Their filenames should have the following format:

      {backend_id}-x86_64-root.dump

    e.g., debian-6.0.1a-x86_64-root.dump (backend_id = "debian-6.0.1a")

14. Setup Synnefo-specific networking on the Ganeti backend:

    This part is deployment-specific and must be customized based on the

    specific needs of the system administrators.

    A reference installation will use a Synnefo-specific KVM ifup script,

    NFDHCPD and pre-provisioned Linux bridges to support public and private

    network functionality. For this:

    Grab NFDHCPD from its own repository (https://code.grnet.gr/git/nfdhcpd),

    install it, modify /etc/nfdhcpd/nfdhcpd.conf to reflect your network

    configuration.

    Install a custom KVM ifup script for use by Ganeti, as

    /etc/ganeti/kvm-vif-bridge, on GANETI-NODEs. A sample implementation is

    provided under /contrib/ganeti-hooks. Set NFDHCPD_STATE_DIR to point

    to NFDHCPD's state directory, usually /var/lib/nfdhcpd.

15. See section "Logging" in README.admin, and edit logic/logging.conf

    according to your OS and individual deployment characteristics.

16. Optionally, read the okeanos_site/README file to setup ~okeanos introductory 

    site (intro, video/info pages). Please see okeanos_site/90-okeanos.sample

    for a sample configuration file which overrides site-specific variables,

    to be placed under settings.d/, after customization.

17. (Hopefully) Done