Didmos consists of different modules (see didmos 2.0). Some of these modules are based on existing open source software with configuration and extensions provided by DAASI (e. g. the LDAP server in didmos Core or didmos Auth). Other modules are developed and shipped by DAASI (e.g. didmos LUI) . Therefore the deployment and operations model for the components is quite heterogenous.
Deployment is supported as either Docker containers or as a VM based deployment for most components with some exceptions as per the following list:
Docker | VM based (Ubuntu) | |
---|---|---|
Core | yes | yes |
LUI | yes | yes |
Authenticator | yes | yes |
Provisioner | yes | no |
ETL | yes | no |
Pwd Synchronizer | no | yes |
For VM based deployment only Ubuntu 18/20 is currently supported.
Docker images are provided for all components (except Pwd Synchronizer) and this is the preferred deployment model.
In order to run didmos as docker containers the following requirements must be met:
- docker: https://docs.docker.com/compose/install/, version 20.10.0 or later
- docker-compose: https://docs.docker.com/compose/install/, version 1.25.0 or later
A docker-compose.yml file describes the system as a whole. See the following documentation and example for didmos2-demo:
https://gitlab.daasi.de/didmos2-demo/didmos2-demo-compose/-/tree/master/deploy
The docker-compose.yml file for individual projects might deviate from this example, as more or less components are included and configuration might be different. Furthermore a .env file must be located in the same directory which contains deployment specific variables.
On the docker host these files are usually located in either /root/docker or /opt/didmos.
The following commands might be useful for operations:
# Start docker-compose up -d # Stop docker-compose down # Display Status docker-compose ps # Show logs of individual container docker logs {container-name} # Restart individual container docker restart {container-name} |
For the VM based deployment project specific Ansible roles are provided for initial setup. The general setup is documented here: https://gitlab.daasi.de/didmos2/didmos2-compose/-/tree/master/ansible
Please note that most didmos projects are extended by project specific roles for setup of extensions and project specific components. Generally these roles are also required for a full setup.
After running the initial setup via Ansible please refer to the following chapters for details on operations for each of the didmos modules:
didmos Core consists of two Docker containers:
Container | Description |
---|---|
{project-name}-core | API |
{project-name}-openldap | LDAP Metadirectory |
The logs of each component can be accessed via docker logs.
docker logs {container-name} |
Configuration is possible via docker environment variables (for supported parameters). An example for this configuration is described in section Provisioner - Docker.
The LDAP Metadirectory is installed via the Ubuntu distribution during the initial Ansible setup (i.e. apt install slapd).
It can be administered using the following commands:
systemctl {start|stop|restart|status} slapd |
slapd logs using rsyslog. The logging is configured in the file rsyslog.conf. The most recent logged messages can be read using this command:
journalctl -r -t openldap |
The configuration of the log level for slapd uses the parameter loglevel in the file /usr/local/etc/openldap/slapd.conf:
... loglevel stats ... |
These log levels are supported:
Numeric values and keywords for log levels are equivalent.
More information about logging and debugging OpenLDAP can be found in the documentation:
The didmos Core API server is installed as a python virtual environment and deployed as a mod_wsgi app in Apache webserver. The following locations on the VM are used:
Component | Location in file system |
---|---|
Python Virtual environment | /opt/didmos2coreEnv |
Python application | /opt/didmos2core |
Configuration (Templates and default config) | /opt/didmos2core/general /opt/didmos2core/customer/customer_config |
Configuration (Overrides) | /etc/didmos/core |
Logs | /var/log/didmos |
Apache config (mod_wsgi integration) | /etc/apache2/sites-available/api-ssl.conf |
Apache logs | /var/log/apache2 |
Restarting the didmos Core API server is possible via the Apache webserver:
systemctl {start|stop|restart|status} apache2 |
The core data of a didmos system is stored in the LDAP server and therefore backups of the entire LDAP server should be done regularly.
This can be done in different ways:
Note that in case of a docker deployment the folders are stored in docker volumes with the following names, which must be backed up:
didmos LUI consists of the following Docker container:
In this container, the compiled frontend (Angular JavaScript app with assets like images, CSS-files etc.) is shipped using an nginx webserver.
The logs can be accessed via docker logs (see list of general commands). Since the application itself runs as Java Script in the web browser, for debugging purposes the browser console might be more useful than the server side logs.
Configuration is possible via docker environment variables (for supported parameters). An example for this configuration is described in section Provisioner - Docker.
The compiled frontend (Angular JavaScript app with assets like images, CSS-files etc.) is located in /var/www/didmos2lui and then shipped as static files using an Apache webserver. The following locations on the VM are used:
Component | Location in file system |
---|---|
Frontend files | /var/www/didmos2lui |
Configuration file | /var/www/didmos2lui/assets/config/environment.json |
Apache config | /etc/apache2/sites-available/lui-ssl.conf |
Apache logs | /var/log/apache2 |
In general, changes to the functionality always require recompiling the static files from source and then redeploying the compiled application on the VM.
didmos Auth consists of the following Docker container:
Container | Description |
---|---|
{project-name}-auth | Application |
{project-name}-mongo | MongoDB Database |
In the -auth container Auth is running as a mod_wsgi application inside an Apache webserver.
The -mongo container is running a MongoDB for storage of the OIDC OP (i.e. registered clients, tokens).
The logs can be accessed via docker logs (see list of general commands).
Configuration is possible via docker environment variables (for supported parameters). For a list of general environment variables refer to didmos2 Authenticator. An example for this configuration is described in section Provisioner - Docker.
didmos Auth is installed as a python virtual environment and deployed as a mod_wsgi app in Apache webserver. The following locations of the VM are used:
Component | Location in file system |
---|---|
Virtual environment (Python) | /opt/didmos2auth |
Configuration | /etc/satosa |
Logs | /var/log/satosa |
Apache config (mod_wsgi integration) | /etc/apache2/sites-available/auth.conf |
Apache logs | /var/log/apache2 |
MongoDB logs | /var/log/mongodb |
Restarting didmos Auth is possible via the Apache webserver:
systemctl {start|stop|restart|status} apache2 |
The application is based on Satosa and most of the configuration in /etc/satosa follows the default Satosa configuration (see https://github.com/IdentityPython/SATOSA).
MongoDB is installed via the Ubuntu distribution during the initial Ansible setup (i.e. apt install mongodb).
The administration of MongoDB can be done using these commands:
systemctl {start|stop|restart|status} mongodb |
Persistent data from MongoDB should be backed up frequently, especially for the registered clients in the OIDC OP. Otherwise they have to be registered again in case of data loss.
Refer to https://docs.mongodb.com/manual/core/backups/ for general backup strategies.
didmos Provisioner consists of the following Docker containers:
Container | Description |
---|---|
didmos2-rabbitmq | RabbitMQ queue |
{project-name}-ra | Requesting authority |
{project-name}-xyz-worker | Worker nodes, possible multiple containers for each target system |
The logs can be accessed via docker logs (see list of general commands).
The composition of the container and basic configuration is done in the file docker-compose.yml.
ra: image: ${CUSTOMER_RA_IMAGE}:${CUSTOMER_RA_TAG} container_name: ${CUSTOMER}-ra depends_on: - rabbitmq volumes: - didmos2-ra-config:/opt/daasi/didmos2/ra/config/ environment: LDAP_URL: ${LDAP_URL} RA_LDAP_ACCOUNT_PW: ${RA_LDAP_ACCOUNT_PW} RA_SENDER_EMAIL: ${RA_SENDER_EMAIL} RA_DEFAULT_RECEIVER_EMAIL: ${RA_DEFAULT_RECEIVER_EMAIL} INIT_PARAM: -i 2 RABBITMQ_URL: ${RABBITMQ_URL} RABBITMQ_USER: ${RABBITMQ_USER} RABBITMQ_PW: ${RABBITMQ_PW} LDAP_ACCESSLOG_PW: ${LDAP_ACCESSLOG_PW} LDAP_MANAGER_PW: ${LDAP_MANAGER_PW} SMTP_HOST: ${SMTP_HOST} SMTP_USER: ${SMTP_USER} SMTP_PASSWORD: ${SMTP_PASSWORD} SMTP_PORT: ${SMTP_PORT} extra_hosts: - "host.docker.internal:host-gateway" |
Additional configuration is possible via docker environment variables (for supported parameters). These variables are referenced in the docker-compose.yml and set in the .env file:
... LDAP_ACCESSLOG_PW=... LDAP_MANAGER_PW=... LDAP_URL=ldap://host.docker.internal:389 LDAP_BIND_DN="cn=manager,dc=didmos,dc=de" RA_LDAP_ACCOUNT_PW=... RA_SENDER_EMAIL=noreply@daasi.de RA_DEFAULT_RECEIVER_EMAIL=noreply@example.com RABBITMQ_URL=rabbitmq RABBITMQ_USER=admin RABBITMQ_PW=... RABBITMQ_PORT=5672 ... |
For a list of general environment variables refer to didmos2 Provisioner.
After changing one or more values in the .env file a recompose has to be performed by using this command:
docker-compose up -d |
A simple restart of the affected containers does not suffice.
The basic principle of docker-compose and docker environment variables applies to all docker deployments for didmos 2.
didmos ETL consists of the following Docker container:
The data and config is mounted as docker volumes from the host system like so (the variables are defined in .env):
volumes: - /${ETL_DATA_DIR}/:/var/didmos/:rw - /${ETL_CONF_DIR}/:/etc/didmos/etl/:rw |
Typically the data and config directories are set as following on the host system:
The logs can be accessed via docker logs (see list of general commands).
Configuration is possible via docker environment variables (for supported parameters) but generally via files in the volume. An example for the configuration using docker environment variables is described in section Provisioner - Docker.