Nosana running jobs at 0: spare GPU or an offline host
A successful 0 is labeled no running jobs. A failed read stays unknown; do not record it as 0. Neither result is hashrate, and neither proves the machine is powered off.
| What you see | Meaning | Next step |
|---|---|---|
| No running jobs, and the number is 0 | This running-list total is 0 | Check online state and eligibility on the official node page |
| Unavailable or a dash | This read did not return a count | Retry the same node address; do not write a blank as 0 |
| Idle GPU on the website | An invitation to offer spare GPUs to jobs | Not the monitor’s 0, and not mining hashrate |
Zero answers “are any jobs running now?”
On the Nosana monitor, running jobs for a node is the task-list total for that address in the running state. A successful 0 sets the status line to no running jobs. Completed jobs can be greater than zero at the same time: a past completion count does not mean a job is running now. How to read both cells is in node counts.
The Nosana site describes idle GPU as spare cards you can share: unused capacity offered to workloads. That is a product description of supply. It is not the definition of this cell. There is no proof-of-work hashrate here. A job is compute work on a GPU, not a hash rate.
What offline would look like
- The monitor does not return online, offline, or a heartbeat. Zero does not prove the host is powered off, disconnected, or failing a driver.
- A failed read leaves both cells unknown. Keep unknown separate from zero. If a later refresh becomes 0, that later read is a successful empty list.
- A network running-job total above zero only means the public index has someone else’s running jobs. It does not put your host online.
A schematic case
Illustrative, not measured income or proof of uptime: the same node shows running 0 and completed 4, with the status no running jobs. An hour later it is still running 0 and completed 4. Those two successful reads only say the running list was empty both times and the completion count did not change. They do not measure whether the GPU had power.
What to do next
- Open this node on the Nosana dashboard and see whether the official page marks the host online, eligible, or assigned work.
- Host GPUs and Troubleshooting live in the Nosana docs. Eligibility, the system, and common failures belong there. Do not substitute a zero.
- If you may have pasted the wrong machine, check node address and wallet before you edit the saved device.
Limits
The monitor cannot reboot the host and cannot read local Nosana logs. Crash analysis does not cover Nosana. The Mac agent controls IOTA Train at Home only.
Common questions
Completed stays at 4 and running stays at 0. Is the host broken?
The two cells only say the list has no new running job and the completion count did not change. Hardware, network, and eligibility are on the official node page.
The site says sharing idle GPUs earns money. Is that this zero?
No. The site is talking about offering spare GPUs to workloads. The monitor’s 0 is an empty running-job list. Neither produces a NOS amount.
The status says unavailable and the figure is a dash. Can I record 0?
No. Unavailable or stale is unknown. Record a successful empty list only when the status is no running jobs and the running cell is 0. Do not record a lookup that is still spinning.