New updates available

openmediavault-clamav 7.0.3

  • Issue #1961: Fix logrotate configuration issue.

openmediavault-k8s 7.4.7

  • Add button to display the recipe README.

owntone and paperless-ngx are available as openmediavault-k8s recipe

Two new recipes for the owntone and paperless-ngx apps have been released to the openmediavault-k8s-recipes GitHub repository. You can get them after updating the recipes list in your openmediavault-k8s plugin.

The following list of applications is now available:

  • filebrowser
  • immich
  • jellyfin
  • minio
  • owntone
  • paperless-ngx
  • photoprism
  • plex-media-server
  • wetty

Contributions from the community are welcome to increase the number of recipes.

Make use of the DSL in the openmediavault-k8s plugin

The recipe editor of the openmediavault-k8s has a DSL (Domain Specific Language) that supports the user in getting specific information from their openmediavault host system in Kubernetes manifests.

The following functions are currently available:

  • hostname() – Get the hostname of the host.
  • domain() – Get the domain name of the host.
  • fqdn() – Get the FQDN of the host.
  • ipaddr(name=NULL) – Get the IPv4 address of the specified network interface. If not set, the IPv4 address of the interface of the first default route is used.
  • ipaddr6(name=NULL) – Get the IPv6 address of the specified network interface. If not set, the IPv6 address of the interface of the first default route is used.
  • uid(name) – Get the UID of the specified user.
  • gid(name) – Get the GID of the specified group.
  • sharedfolder_path(name) – Get the full path of the specified shared folder.
  • conf_get(id, uuid=NULL) – Get the specified database configuration object.
  • tz() – Get the time zone of the host.

The following filters are currently available:

  • get(key, default=NULL) – Get the specified key, e.g. foo.bar.baz from a dictionary.

For the built-in features of the used template engine please have a look here:

Here are several examples that make use of the DSL:

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app.kubernetes.io/instance: wetty
    app.kubernetes.io/name: wetty
  name: wetty
  namespace: wetty-app
spec:
  ...
  template:
    ...
    spec:
      containers:
        - image: docker.io/wettyoss/wetty:latest
          imagePullPolicy: IfNotPresent
          name: wetty
          ports:
            - containerPort: 3000
          securityContext:
            runAsNonRoot: true
            runAsUser: {{ uid('nobody') }}
            runAsGroup: {{ gid('nogroup') }}
          args:
            - "--port=3000"
            - "--base=/"
            - "--force-ssh"
            - "--ssh-port={{ conf_get('conf.service.ssh') | get('port') }}"
            - "--ssh-host={{ hostname() }}"
      restartPolicy: Always
...
persistence:
    import:
        enabled: true
        type: hostPath
        readOnly: true
        mountPath: /srv/import
        hostPath: {{ sharedfolder_path('images') }}
        hostPathType: Directory
apiVersion: v1
kind: PersistentVolume
metadata:
  name: config-dir
  labels:
    app.kubernetes.io/instance: plex-media-server
    app.kubernetes.io/name: plex-media-server
spec:
  storageClassName: shared-folder
  capacity:
    storage: 2Gi
  hostPath:
    path: {{ sharedfolder_path('media') }}
    type: Directory
  accessModes:
    - ReadWriteMany
apiVersion: traefik.containo.us/v1alpha1
kind: IngressRoute
metadata:
  name: immich-websecure
  namespace: immich-app
  ...
spec:
  entryPoints:
    - websecure
  routes:
    - match: Host(`immich.{{ ipaddr() }}.sslip.io`)
      kind: Rule
      services:
        - name: immich-server
          port: 3001

Problems with DHCP network configuration

We encountered a strange bug in systemd-networkd that causes some troubles on systems which are using DHCP to configure the network interfaces. Because of unknown reasons the systemd-networkd service may crash, e.g. when a new network configuration is applied.

You can workaround by rebooting the system or by running

systemctl restart systemd-networkd.service

several times until

systemctl status systemd-networkd.service

reports the service is running. You can alternatively run the command

omv-salt deploy run systemd-networkd

If everything fails, use static IP addresses until the bug has been fixed upstream by the Debian project.

The bug has been reported upstream at https://github.com/systemd/systemd/issues/27854 and https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1036949.

What is wrong with some Open-Source-Software (OSS) users?

A few days ago I received another email which made me wonder what is going on with some Open-Source-Software (OSS) users.

… Try fresh install as I described to you and do your job and fix this ridiculous bug.

First of all, the writer here has quite the wrong tone. Secondly, I have already done my job by extensively testing the functionality before publishing. Of course, it always happens that things run differently at the user than in the development or test environment.

What always saddens me is the fact that some users think they can make demands on software that is created by developers in their spare time and then provided to them for free. How can you make claims on a software and its developers that is free? With paid software I can understand that, but with OSS? In my opinion, many users do not see the difference between commercially driven OSS and those created by volunteers in their spare time. Projects like Linux, Kubernetes, Ceph or OpenOffice, to name a few, are supported by vendors who have a commercial interest and have paid developers working on these projects for this purpose. Such projects have naturally set the bar pretty high for users. It is clear that many users then apply this bar to other OSS projects. However, users should not forget how open source works and where it comes from. In most cases, OSS is driven by enthusiastic developers in their spare time and made available to the community for free.