Alert.png The wiki is deprecated and due to be decommissioned by the end of September 2022.
The content is being migrated to other supports, new updates will be ignored and lost.
If needed you can get in touch with EGI SDIS team using operations @ egi.eu.

Federated Cloud Mitaka guide

From EGIWiki
Jump to navigation Jump to search
Overview For users For resource providers Infrastructure status Site-specific configuration Architecture



Introduction

Integration with FedCloud requires a working OpenStack installation as a prerequirement (see http://docs.openstack.org/ for details). This manual provides information on how to set up a Resource Centre providing cloud resources in the EGI infrastructure, using OpenStack Mitaka. FedCloud components are distributed als through CMD (Cloud Middleware Distribution); for components distributed with CMD, corresponding instructions will be provided.

EGI currently operates two realms:

  • the Open Standards Realm and
  • the OpenStack Realm

Both are completely integrated with the EGI federation services but they expose different interfaces to offer IaaS capabilities to the users. The Open Standards Realm uses OCCI standard (supported by providers with OpenNebula, OpenStack and Synnefo cloud management frameworks), while the OpenStack Realm uses the OpenStack native Nova API (support limited to OpenStack providers).

You can provide your service in one or both of the realms. For the OpenStack Realm, you just need to declare your endpoint in GOCDB as described below. For the Open Standards realm you will need to deploy an additional service for providing OCCI access.

General minimal requirements:

  • very minimal hardware required to join. Hardware requirements depend on:
    • the cloud stack you use
    • the amount of resources you want to make available
    • the number of users/use cases you want to support
  • Servers need to authenticate each other in the EGI Federated Cloud context; this is fulfilled using X.509 certificates, so a Resource Centre should be able to obtain server certificates for some services.
  • User and research communities are called Virtual Organisations (VO). At least support for 2 VOs is needed to join as a Resource Centre:
    • ops and dteam, used for operational purposes as per RC OLA
    • Most providers also support fedcloud.egi.eu VO for application prototyping and validation.
  • The operating systems supported by the UMD are:
    • CentOS7
    • Ubuntu Trusty/Xenial

Components description

Which components must be installed and configured depends on the services the RC wants to provide.

  • Keystone (the authorization component) must always be provided
  • If VM Management features are offered (OCCI access or OpenStack access), then Nova, Cinder and Glance must be available
  • also Neutron is needed, but nova-network can also be used for legacy installations (see here for instructions on how to migrate to Neutron).

Openstack-fedcloud.png

As you can see from the schema above, the integration is performed installing some EGI extensions on top of the OpenStack components.

  • Keystone-VOMS Authorization plugin allow users with a valid VOMS proxy to access the OpenStack deployment
  • OpenStack OCCI Interface (ooi) translates between OpenStack API and OCCI
  • cASO collecting accounting data from OpenStack + SSM sending the records extracted by cASO to the central accounting database on the EGI Accounting service (APEL)
  • BDII cloud provider registers the RC configuration and description through the EGI Information System to facilitate service discovery
  • vmcatcher checks the EGI App DB for new or updated images that can be provided by the RC to the user communities (VO) supported
  • The vmcatcher hooks (glancepush and OpenStack handler for vmcatcher) push updated subscribed images from vmcatcher to Glance, using Openstack Python API

Configuring the EGI User Management/AAI (Keystone-VOMS)

Every FedCloud site must support authentication of users with X.509 certificates with VOMS extensions. The Keystone-VOMS extension enables this kind of authentication on Keystone.

Documentation on the installation is available on https://keystone-voms.readthedocs.io/en/stable-mitaka/

Notes:

  • You need a host certificate from a recognised CA for your keystone server.
  • Take into account that using keystone-voms plugin will enforce the use of https for your Keystone service, you will need to update your URLs at the Keystone catalog and in the configuration of your services:
    • You will probably need to include your CA to your system's CA bundle to avoid certificate validation issues: /etc/ssl/certs/ca-certificates.crt from the ca-certificates package on Debian/Ubuntu systems or /etc/pki/tls/certs/ca-bundle.crt from the ca-certificates on RH and derived systems. The Federated Cloud OpenStack Client guide includes information on how to do it.
    • replace http with https in auth_[protocol|uri|url] and auth_[host|uri|url] in the nova, cinder, glance and neutron config files (/etc/nova/nova.conf, /etc/nova/api-paste.ini, /etc/neutron/neutron.conf, /etc/neutron/api-paste.ini, /etc/neutron/metadata_agent.ini, /etc/cinder/cinder.conf, /etc/cinder/api-paste.ini, /etc/glance/glance-api.conf, /etc/glance/glance-registry.conf, /etc/glance/glance-cache.conf) and any other service that needs to check keystone tokens.
    • You can update the URLs of the services directly in the database:
mysql> use keystone;
mysql> update endpoint set url="https://<keystone-host>:5000/v2.0" where url="http://<keystone-host>:5000/v2.0";
mysql> update endpoint set url="https://<keystone-host>:35357/v2.0" where url="http://<keystone-host>:35357/v2.0";
  • Support for EGI VOs: VOMS configuration, you should configure fedcloud.egi.eu, dteam and ops VOs.
  • VOMS-Keystone configuration: most sites should enable the autocreate_users option in the [voms] section of Keystone-VOMS configuration. This will enable that new users are automatically created in your local keystone the first time they login into your site.

Configuring the EGI VM Management Interface (OCCI interface)

ooi is the recommended software to provide OCCI for OpenStack. Installation and configuration of ooi is available at ooi documentation. Packages currently supported and working on Mitaka can be found at https://appdb.egi.eu/store/software/ooi/releases/occi-1.1/

Once the OCCI interface is installed, you should register it on your installation (adapt the region and URL to your deployment):

$ openstack service create --name occi --description "OCCI Interface" occi
+-------------+----------------------------------+
| Field       | Value                            |
+-------------+----------------------------------+
| description | OCCI Interface                   |
| enabled     | True                             |
| id          | 6dfd6a56c9a6456b84e8c86038e58f56 |
| name        | occi                             |
| type        | occi                             |
+-------------+----------------------------------+

$ openstack endpoint create --region RegionOne occi --publicurl http://172.16.4.70:8787/occi1.1

+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |           OCCI service           |
|      id     | 8e6de5d0d7624584bed6bec9bef7c9e0 |
|     name    |             occi_api             |
|     type    |               occi               |
+-------------+----------------------------------+

Configuring Accounting, Information System, Image Management

There are two options to install the remaining Accounting, Information System and Image Management components:

Integration using EGI FedCloud Appliance

The EGI FedCloud Appliance:

https://appdb.egi.eu/store/vappliance/fedcloud.integration.appliance.openstack

packages a set of docker containers to federate a OpenStack deployment with some EGI services:

  • Information System (BDII)
  • Accounting (cASO, SSM)
  • Image management (atrope)

It is available as an OVA file. You can easily extract the VMDK disk of the OVA by untaring the file.

Pre-requisites

The appliance works by querying the public APIs of an existing OpenStack installation. It assumes Keystone-VOMS is installed at that OpenStack and the voms.json file is properly configured.

The appliance uses the following OpenStack APIs:

  • nova, for getting images and flavors available and to get usage information
  • keystone, for authentication and for getting the available tenants
  • glance, for querying, uploading and removing VM images.

Not all services need to be accessed with the same credentials. Each component is individually configured.

A host certificate is to send the accounting information before sending it to the accounting repository. DN of the host certificate must be registered in GOCDB service type eu.egi.cloud.accounting (see the registration section below for more information).

Note:

  • VM Image replication requires large disk space for storing the downloaded images. By default these are stored at /image_data. You can mount a volume at that location.
  • The appliance should be accessible by the EGI Information System. EGI information system will check GOCDB for the exact location of your appliance (see the registration section below for more information).

EGI Accounting (cASO/SSM)

There are two different processes handling the accounting integration:

  • cASO, which connects to the OpenStack deployment to get the usage information, and,
  • ssmsend, which sends that usage information to the central EGI accounting repository.

They are run by cron every hour (cASO) and every six hours (ssmsend).

cASO configuration is stored at /etc/caso/caso.conf. Most default values are ok, but you must set:

  • site_name (line 100)
  • tenants (line 104)
  • credentials to access the accounting data (lines 122-128). Check the cASO documentation for the expected permissions of the user configured here.

The cron job will use the voms mapping file at /etc/voms.json.

cASO will write records to /var/spool/apel where ssmsend will take them.

SSM configuration is available at /etc/apel. Defaults should be ok for most cases. The cron file uses /etc/grid-security for the CAs and the host certificate and private keys (in /etc/grid-security/hostcert.pem and /etc/grid-security/hostkey.pem).

Running the services

Both caso and ssmsend are run via cron scripts. They are located at /etc/cron.d/caso and /etc/crond.d/ssmsend respectively. For convenience there are also two scripts /usr/loca/bin/caso-extract.sh and /usr/local/bin/ssm-send.sh that run the docker container with the proper volumes.

EGI Information System (BDII)

Information discovery provides a real-time view about the actual images and flavors available at the OpenStack for the federation users. It has two components:

  • Resource-Level BDII: which queries the OpenStack deployment to get the information to publish
  • Site-Level BDII: gathers information from several resource-level BDIIs (in this case only 1) and makes it publicly available for the EGI information system.

Resource-level BDII

This is provided by container egifedcloud/cloudbdii. You need to configure:

  • /etc/cloud-info-provider/openstack.rc, with the credentials to query your OpenStack. The user configured just needs to be able to access the lists of images and flavors.
  • /etc/cloud-info-provider/openstack.yaml, this file includes the static information of your deployment. Make sure to set the SITE-NAME as defined in GOCDB.

Site-level BDII

The egifedcloud/sitebdii container runs this process. Configuration files:

  • /etc/sitebdii/glite-info-site-defaults.conf. Set here the name of your site (as defined in GOCDB) and the public hostname where the appliance will be available.
  • /etc/sitebdii/site.cfg. Include here basic information on your site.

Running the services

In order to run the information discovery containers, there is a docker-compose file at /etc/sitebdii/docker-compose.yml. Run it with:

docker-compose -f /etc/sitebdii/docker-compose.yml up -d

Check the status with:

docker-compose -f /etc/sitebdii/docker-compose.yml ps

You should be able to get the BDII information with an LDAP client, e.g.:

ldapsearch -x -p 2170 -h <yourVM.hostname.domain.com> -b o=glue

EGI Image Management (atrope)

The appliance provide VMI replication with atrope, an alternative implementation to vmcatcher. Every 12 hours, the appliance will perform the following actions:

  • download the configured lists in /etc/atrope/hepix.yaml and verify its signature
  • check any changes in the lists and download new images
  • synchronise this information to the configured glance endpoint

Configure the glance credentials in the /etc/atrope/atrope.conf file and add the lists you want to download at the /etc/atrope/hepix.yaml. See the following example for fedcloud.egi.vo list:

# This must match the VO name configured at the voms.json file
fedcloud.egi.eu:
    url: https://vmcaster.appdb.egi.eu/store/vo/fedcloud.egi.eu/image.list
    enabled: true
    # All image lists from AppDB will have this endorser
    endorser:
        dn: '/DC=EU/DC=EGI/C=NL/O=Hosts/O=EGI.eu/CN=appdb.egi.eu'
        ca: "/DC=ORG/DC=SEE-GRID/CN=SEE-GRID CA 2013"
    # You must get this from AppDB
    token: 17580f07-1e33-4a38-94e3-3386daced5be
    # if you want to restrict the images downloaded from the AppDB, you can add here a list of the identifiers
    # check the "dc:identifier" field in the image list file.
    images: []
    # images names will prefixed with this string for easy identification
    prefix: "FEDCLOUD "

Check How to subscribe to a private image list for instructions to get the URL and token. The prefix if specified will be added in the image title in glance. You can define a subset of images to download with the images field.

Running the service

atrope is run via a cron scripts: /etc/cron.d/atrope. For convenience the /usr/loca/bin/atrope-dispatch.sh script runs the docker container with the proper volumes.

Integration using individual components

EGI Accounting (cASO/SSM)

Every cloud RC should publish utilization data to the EGI accounting database. You will need to install cASO, a pluggable extractor of Cloud Accounting Usage Records from OpenStack.

Documentation on how to install and configure cASO is available at https://caso.readthedocs.org/en/latest/

In order to send the records to the accounting database, you will also need to configure SSM, whose documentation can be found at https://github.com/apel/ssm

EGI Information System (BDII)

Sites must publish information to EGI information system which is based on BDII. The BDII can be installed easily directly from the distribution repository, the package is usually named "bdii".

There is a common cloud information provider for all cloud management frameworks that collects the information from the used CMF and send them to the aforementioned BDII. It can be installed on the same machine as the BDII or on another machine. The installation and configuration guide for the cloud information provider can be found in the following Fedclouds BDII instructions; more detailed installation and configuration instructions are available at https://github.com/EGI-FCTF/cloud-bdii-provider

EGI Image Management (vmcatcher, glancepush)

Sites in FedCloud offering VM management capability must give access to VO-endorsed VM images. This functionality is provided with vmcatcher (that is able to subscribe to the image lists available in AppDB) and a set of tools that are able to push the subscribed images into the glance catalog. In order to subscribe to VO-wide image lists, you need to have a valid access token to the AppDB. Check how to access to VO-wide image lists and how to subscribe to a private image list documentation for more information.

Please refer to vmcatcher documentation for installation.

Vmcatcher may be branched to Openstack Glance catalog using python-glancepush tool and Openstack Handler for Vmcatcher event handler. To install and configure glancepush and the handler, you can refer to the following instructions:

[stack@ubuntu]$ wget http://repository.egi.eu/community/software/python.glancepush/0.0.X/releases/generic/0.0.6/python-glancepush-0.0.6.tar.gz
[stack@ubuntu]$ tar -zxvf python-glancepush-0.0.6.tar.gz
[stack@ubuntu]$ python setup.py install
    • for RHEL6 you can run:
[stack@rhel]$ yum localinstall http://repository.egi.eu/community/software/python.glancepush/0.0.X/releases/sl/6/x86_64/RPMS/python-glancepush-0.0.6-1.noarch.rpm
  • Then, configure glancepush directories
[stack@ubuntu]$ sudo mkdir -p /var/spool/glancepush /etc/glancepush/log /etc/glancepush/transform/ /etc/glancepush/clouds /var/log/glancepush
[stack@ubuntu]$ sudo chown stack:stack -R /var/spool/glancepush /etc/glancepush /var/log/glancepush/
  • Copy the file /etc/keystone/voms.json to /etc/glancepush/voms.json. Then create a file in clouds file for every VO to which you are subscribed. For example, if you're subscribed to fedcloud, atlas and lhcb, you'll need 3 files in the /etc/glancepush/clouds directory with the credentials for this VO/tenants, for example:
[general]
# Tenant for this VO. Must match the tenant defined in voms.json file
testing_tenant=egi
# Identity service endpoint (Keystone)
endpoint_url=https://server4-eupt.unizar.es:5000/v2.0
# User Password
password=123456
# User
username=John
# Set this to true if you're NOT using self-signed certificates
is_secure=True
# SSH private key that will be used to perform policy checks (to be done)
ssh_key=Carlos_lxbifi81
# WARNING: Only define the next variable if you're going to need it. Otherwise you may encounter problems
cacert=path_to_your_cert
[stack@ubuntu]$ wget http://repository.egi.eu/community/software/openstack.handler.for.vmcatcher/0.0.X/releases/generic/0.0.7/gpvcmupdate-0.0.7.tar.gz
[stack@ubuntu]$ tar -zxvf gpvcmupdate-0.0.7.tar.gz
[stack@ubuntu]$ python setup.py install

while for RHEL6 you can run:

[stack@rhel]$ yum localinstall http://repository.egi.eu/community/software/openstack.handler.for.vmcatcher/0.0.X/releases/sl/6/x86_64/RPMS/gpvcmupdate-0.0.7-1.noarch.rpm
  • Create the vmcatcher folders for OpenStack
[stack@ubuntu]$ mkdir -p /opt/stack/vmcatcher/cache /opt/stack/vmcatcher/cache/partial /opt/stack/vmcatcher/cache/expired
  • Check that vmcatcher is running properly by listing and subscribing to an image list
[stack@ubuntu]$ export VMCATCHER_RDBMS="sqlite:////opt/stack/vmcatcher/vmcatcher.db"
[stack@ubuntu]$ vmcatcher_subscribe -l
[stack@ubuntu]$ vmcatcher_subscribe -e -s https://vmcaster.appdb.egi.eu/store/vappliance/tinycorelinux/image.list
[stack@ubuntu]$ vmcatcher_subscribe -l
8ddbd4f6-fb95-4917-b105-c89b5df99dda    True    None    https://vmcaster.appdb.egi.eu/store/vappliance/tinycorelinux/image.list
  • Create a CRON wrapper for vmcatcher, named $HOME/gpvcmupdate/vmcatcher_eventHndl_OS_cron.sh, using the following code
#!/bin/bash
#Cron handler for VMCatcher image syncronization script for OpenStack

#Vmcatcher configuration variables
export VMCATCHER_RDBMS="sqlite:////opt/stack/vmcatcher/vmcatcher.db"
export VMCATCHER_CACHE_DIR_CACHE="/opt/stack/vmcatcher/cache"
export VMCATCHER_CACHE_DIR_DOWNLOAD="/opt/stack/vmcatcher/cache/partial"
export VMCATCHER_CACHE_DIR_EXPIRE="/opt/stack/vmcatcher/cache/expired"
export VMCATCHER_CACHE_EVENT="python $HOME/gpvcmupdate/gpvcmupdate.py -D"

#Update vmcatcher image lists
vmcatcher_subscribe -U

#Add all the new images to the cache
for a in `vmcatcher_image -l | awk '{if ($2==2) print $1}'`; do
 vmcatcher_image -a -u $a
done 

#Update the cache
vmcatcher_cache -v -v

#Run glancepush
/usr/bin/glancepush.py
  • Set the newly created file as executable
[stack@ubuntu]$ chmod +x $HOME/gpvcmupdate/vmcatcher_eventHndl_OS_cron.sh
  • Test that the vmcatcher handler is working correctly by running
[stack@ubuntu]$ $HOME/gpvcmupdate/vmcatcher_eventHndl_OS_cron.sh
INFO:main:Defaulting actions as 'expire', and 'download'.
DEBUG:Events:event 'ProcessPrefix' executed 'python /opt/stack/gpvcmupdate/gpvcmupdate.py'
DEBUG:Events:stdout=
DEBUG:Events:stderr=Ignoring ProcessPrefix event.
INFO:DownloadDir:Downloading '541b01a8-94bd-4545-83a8-6ea07209b440'.
DEBUG:Events:event 'AvailablePrefix' executed 'python /opt/stack/gpvcmupdate/gpvcmupdate.py'
DEBUG:Events:stdout=AvailablePrefix
DEBUG:Events:stderr=
INFO:CacheMan:moved file 541b01a8-94bd-4545-83a8-6ea07209b440
DEBUG:Events:event 'AvailablePostfix' executed 'python /opt/stack/gpvcmupdate/gpvcmupdate.py'
DEBUG:Events:stdout=AvailablePostfixCreating Metadata Files
DEBUG:Events:stderr=
DEBUG:Events:event 'ProcessPostfix' executed 'python /opt/stack/gpvcmupdate/gpvcmupdate.py'
DEBUG:Events:stdout=
DEBUG:Events:stderr=Ignoring ProcessPostfix event.


  • Add the following line to the stack user crontab:
50 */6 * * * $HOME/gpvcmupdate/vmcatcher_eventHndl_OS_cron.sh >> /var/log/glancepush/vmcatcher.log 2>&1

NOTES:

  • It is recommended to execute glancepush and vmcatcher_cache as stack or other non-root user.
  • VMcatcher expired images are removed from OS.

Post-installation

After the installation of all the needed components, it is recommended to set the following policies on Nova to avoid users accessing other users resources:

[root@egi-cloud]# sed -i 's|"admin_or_owner":  "is_admin:True or project_id:%(project_id)s",|"admin_or_owner":  "is_admin:True or project_id:%(project_id)s",\n    "admin_or_user":  "is_admin:True or user_id:%(user_id)s",|g' /etc/nova/policy.json
[root@egi-cloud]# sed -i 's|"default": "rule:admin_or_owner",|"default": "rule:admin_or_user",|g' /etc/nova/policy.json
[root@egi-cloud]# sed -i 's|"compute:get_all": "",|"compute:get": "rule:admin_or_owner",\n    "compute:get_all": "",|g' /etc/nova/policy.json

Registration, validation and certification

As mentioned in the main page, RC services must be registered in the EGI Configuration Management Database (GOCDB). If you are creating a new site for your cloud services, please follow the Resource Centre Registration and Certification with the help of EGI Operations and your reference Resource Infrastructure.

You will need to register the following services (all of them can be provided by the Federated Cloud Appliance):

  • Site-BDII. This service collects and publishes site's data for the Information System. Existing sites should already have this registered.
  • eu.egi.cloud.accounting. Register here the host sending the records to the accounting repository (executing SSM send).
  • eu.egi.cloud.vm-metadata.vmcatcher for the VMI replication mechanism. Register here the host providing the replication.

If offering OCCI interface, the site must register also:

  • eu.egi.cloud.vm-management.occi for the OCCI endpoint offered by the site. Please note the special endpoint URL syntax described at GOCDB usage in FedCloud

If offering native OpenStack access, you must register:

  • org.openstack.nova for the Nova endpoint of the site. Please note the special endpoint URL syntax described at GOCDB usage in FedCloud


Site should also declare the following properties using the Site Extension Properties feature:

    1. Max number of virtual cores for VM with parameter name: cloud_max_cores4VM
    2. Max amount of RAM for VM with parameter name: cloud_max_RAM4VM using the format: value+unit, e.g. "16GB".
    3. Max amount of storage that could be mounted in a VM with parameter name: cloud_max_storage4VM using the format: value+unit, e.g. "16GB".

The installation validation is part of the aforementioned Resource Centre Registration and Certification procedure. After you register the services in GOCDB, EGI Operations will test your services using the site certification manual tests mentioned in the same procedure. It is important to use that guide to test the services published to check that they are behaving properly.

Once the site services are registered in GOCDB (and flagged as "monitored") they will appear in the EGI service monitoring tools. EGI will check the status of the services (see Infrastructure Status for details). Check if your services are present in the EGI service monitoring tools and passing the tests; if you experience any issues (services not shown, services are not OK...) please contact back EGI Operations or your reference Resource Infrastructure.