Service and processing

Commands in this group control the live pipeline on the appliance guest: systemd units, TSDuck ingress/egress, ordered stack start/stop, offline file processing, digest-pinned container refresh, and NVENC helper workflows. Use them after networking and licensing are sane — they can restart services, apply TSDuck configuration, or pull new container images.

Typical workflow

When SRT is connected but enhancement looks wrong, separate service health from stream health. Start with read-only checks in Health and diagnostics (pixop-version, pixop-preflight, pixop-healthcheck). If units are down or flapping, work here:

pixop-stack status
pixopctl status
tsduckctl check main
tsduckctl check egress-main

To bring the stack up in dependency order, use pixop-stack. For day-two restarts of the Pixop LIVE unit alone, use pixopctl. Ingress and egress TSDuck instances are managed with tsduckctl — credentials, PID detection, latency tuning, and logs.

Offline transcoding or batch enhancement on files already on the appliance uses pixop-process. When Pixop publishes a new digest-pinned LIVE image for your release line, pixop-image-update applies it; it does not upgrade the host OS or NVIDIA driver. Encoder-side experiments and NVENC-focused diagnostics are covered in pixop-nvencc.

Mutating vs read-only

Guest tools under /opt/pixop/bin self-elevate with sudo when required — do not wrap them in extra sudo unless a page explicitly says otherwise.

Release line caveats

Marketplace and KVM deliveries can ship different CLI surfaces on different schedules. If a command is missing, check pixop-version, command -v <tool>, and your procurement record before assuming documentation is wrong.

Command pages

Back to the full CLI reference.