OMD "Labs" Edition (https://labs.consol.de/de/omd/) on Docker with Ansible support.
Author: Sven Nierlein, sven.nierlein at consol.de
Original Author: Simon Meggle
Each image build gets triggered by the OMD Labs build system as soon as there are new packages of OMD available:
- https://hub.docker.com/r/consol/omd-labs-rocky/ (x86 / arm64)
- https://hub.docker.com/r/consol/omd-labs-debian/ (x86 / arm64)
Automated builds are triggered for the following branches:
- master => :nightly (=snapshot builds)
- vX.XX => :vX.XX (=stable version)
- latest => :latest (=latest stable version)
Each image already contains a "demo" site.
Run the "demo" site in OMD Labs Edition:
# Rocky 9
docker run -p 8443:443 consol/omd-labs-rocky
# Debian 11
docker run -p 8443:443 consol/omd-labs-debian
Use the Makefile to work with locally built images:
# run a local image
make -f Makefile.omd-labs-rocky start
# build a "local/" image without overwriting the consol/ image
make -f Makefile.omd-labs-rocky build
# start just the bash
make -f Makefile.omd-labs-rocky bash
The container will log its startup process:
Config and start OMD site: demo
--------------------------------------
Checking for volume mounts...
--------------------------------------
* local/: [No Volume]
* etc/: [No Volume]
* var/: [No Volume]
Checking for Ansible drop-in...
--------------------------------------
Nothing to do (/root/ansible_dropin/playbook.yml not found).
omd-labs: Starting site demo...
--------------------------------------
Preparing tmp directory /omd/sites/demo/tmp...Starting rrdcached...OK
Starting npcd...OK
Starting naemon...OK
Starting dedicated Apache for site demo...OK
Initializing Crontab...OK
OK
Notice the section "Data volume check". In this case there were no host mounted data volumes used. The "start.sh" script has renamed all .ORIG folders to the original name in case there are no mounted volumes.
The default sitename "demo" can be changed. Build a custom image while SITENAME is set:
- clone this repository,
cdinto the folder containg the Dockerfile, e.g.omd-labs-rocky - build a local image:
export SITENAME=mynewsite; make -f Makefile.omd-labs-rocky build
Each container instance of this image will start the site "mynewsite" instead of "demo".
If you need to set the sitename dynamically without building the whole image from scratch (see above), you can create images with another OMD-site beneath the default site (see above). Create a custom Dockerfile which uses the original image as the base image:
FROM: consol/omd-labs-rocky:nightly
...
...
The sitename in your custom OMD image can now be changed by setting the variable NEW_SITENAME to a new value:
export NEW_SITENAME=anothersite
The ONBUILD commands in the original Dockerfile execute after the current Dockerfile build completes. ONBUILD executes in any child image derived FROM the current image. Think of the ONBUILD command as an instruction the parent Dockerfile gives to the child Dockerfile.
As soon as the container dies, all monitoring data (configuration files, RRD data, InfluxDB, log files etc.) are lost, too. To keep the data persistent, use host mounted volumes.
This command
make -f Makefile.omd-labs-rocky startvol
starts the container with three volume mounts:
./site/etc=>$OMD_ROOT/etc.mount./site/local=>$OMD_ROOT/local.mount./site/var=>$OMD_ROOT/var.mount
On the very first start, this folders will be created on the host file system.
In that case, the start.sh synchronize ongoing through lsycnd the content into the volumes (etc.mount, local.mount, var.mount) from the original folders (etc, local, var):
$OMD_ROOT/etc=>$OMD_ROOT/etc.mount$OMD_ROOT/local=>$OMD_ROOT/local.mount$OMD_ROOT/var=>$OMD_ROOT/var.mount
Config and start OMD site: demo
--------------------------------------
Checking for volume mounts...
--------------------------------------
* local/: [EXTERNAL Volume] at /opt/omd/sites/demo/local.mount
* mounted volume is writable
=> local.mount is empty; initial sync from local local ...
* writing the lsyncd config for local.mount...
* etc/: [EXTERNAL Volume] at /opt/omd/sites/demo/etc.mount
* mounted volume is writable
=> etc.mount is empty; initial sync from local etc ...
* writing the lsyncd config for etc.mount...
* var/: [EXTERNAL Volume] at /opt/omd/sites/demo/var.mount
* mounted volume is writable
=> var.mount is empty; initial sync from local var ...
* writing the lsyncd config for var.mount...
lsyncd: Starting lsyncd ...
--------------------------------------
16:38:44 Normal: --- Startup, daemonizing ---
16:38:44 Normal: --- Startup, daemonizing ---
Checking for Ansible drop-in...
--------------------------------------
Nothing to do (/root/ansible_dropin/playbook.yml not found).
omd-labs: Starting site demo...
--------------------------------------
Preparing tmp directory /omd/sites/demo/tmp...Starting rrdcached...OK
Starting npcd...OK
Starting naemon...OK
Starting dedicated Apache for site demo...OK
Initializing Crontab...OK
OK
On the next start the folders are not empty anymore and used as usual.
Before OMD starts, each of the three mount points (etc, local, var) can be checked for free available disk space to ensure that the container can store its data. The threshold can be given as an environent variable in the docker run command. If there is not enough space on any mount point, the startup script fails. On a container orchestration platform like OpenShift this should be handled by your deployment config (=the running pod only gets shutdown if the new pod was started properly).
docker run -d -p 8443:443 \
-v $(pwd)/site/local:/omd/sites/demo/local.mount \
-v $(pwd)/site/etc:/omd/sites/demo/etc.mount \
-v $(pwd)/site/var:/omd/sites/demo/var.mount \
-e VOL_VAR_MB_MIN=700000 \
-e VOL_ETC_MB_MIN=500 \
-e VOL_LOCAL_MB_MIN=6000 \
consol/omd-labs-rocky:nightly
docker logs 91992828cc1dca7839cb2842933897b94329fe2c6b395c5ccb8b9fa056057679
Config and start OMD site: demo
--------------------------------------
Checking for volume mounts...
--------------------------------------
* local/: [EXTERNAL Volume] at /opt/omd/sites/demo/local.mount
* OK: Free space on /opt/omd/sites/demo/local.mount is 499826MB (required: 6000MB)
* OK: mounted volume is writable
<= Volume contains data; sync into local local ...
* writing the lsyncd config for local.mount...
* etc/: [EXTERNAL Volume] at /opt/omd/sites/demo/etc.mount
* OK: Free space on /opt/omd/sites/demo/etc.mount is 499825MB (required: 500MB)
* OK: mounted volume is writable
<= Volume contains data; sync into local etc ...
* writing the lsyncd config for etc.mount...
* var/: [EXTERNAL Volume] at /opt/omd/sites/demo/var.mount
* ERROR: Mounted volume has only 499825MB left (required: 700000MB, set by VOL_VAR_MB_MIN).
no crontab for demo
Removing Crontab...Stopping Nagflux.... Not running.
Stopping dedicated Apache for site demo...(not running)...OK
Stopping naemon...not running...OK
Stopping Grafana.... Not running.
Stopping influxdb.... Not running.
To test if everything worked, simply start the container with
make startvol
This starts the container with the three data volumes. Everything the container writes into one of those three folder, it will synchronized into the persistent file system.
(make startvol is just a handy shortcut to bring up the container. In Kubernetes/OpenShift you won't need this.)
For some time OMD-Labs comes with full Ansible support, which we can use to modify the container instance on startup. How does this work?
By default, the OMD-labs containers start with the CMD /root/start.sh. This script
- checks if there is a
playbook.ymlin$ANSIBLE_DROPIN(default:/root/ansible_dropin, changeable by environment). If found, the playbook is executed. It is completely up to you if you only place one single task inplaybook.yml, or if you also include Ansible roles. (with a certain point of complexity, you should think about a separate image, though...) - starts the OMD site "demo" & Apache as a foreground process
Just a folder containing a valid playbook into the container:
docker run -it -p 8443:443 -v $(pwd)/my_ansible_dropin:/root/ansible_drop consol/omd-labs-debian
When starting the container, OMD will create a random default password for the omdadmin user. There are several ways to handle this:
- using the data volume will bring your own htpasswd file
- set your default omdadmin password per ansbible dropin, ex. like:
playbook.yml:
---
- hosts: all
tasks:
- shell: sudo su - demo -c "set_admin_password omd"
If you want to see more verbose output from Ansible to debug your role, adjust the environment variable value ANSIBLE_VERBOSITY to e.g. 3:
docker run -it -p 8443:443 -e ANSIBLE_VERBOSITY=3 -v $(pwd)/my_ansible_dropin:/root/ansible_drop consol/omd-labs-debian