> Section: [7. Running it](https://jaque.sh/docs/deployment/choose.md)
> Next: deployment/guides/systemd-single
> Index: https://jaque.sh/llms.txt


Two questions decide which guide to follow: where does jaque run, and how
many nodes does it run on. [Topologies](https://jaque.sh/docs/deployment/topologies.md) is the mechanism
behind the answers; this page is the shortcut to a guide that walks you
from a fresh machine to a running jaque, one command at a time.

| | One node | Several nodes |
|---|---|---|
| **systemd host** (deb/rpm) | [Single node, systemd](https://jaque.sh/docs/deployment/guides/systemd-single.md) -- one host, the packaged unit, nothing external. | [Split by role, systemd](https://jaque.sh/docs/deployment/guides/systemd-distributed.md) -- one unit per role across hosts you manage yourself. |
| **Docker Compose** | [Single node, compose](https://jaque.sh/docs/deployment/guides/compose-single.md) -- one container, a named volume for the event log. | [Split by role, compose](https://jaque.sh/docs/deployment/guides/compose-distributed.md) -- one container per role plus a log server, all in one stack. |
| **Kubernetes** | [Single node, Kubernetes](https://jaque.sh/docs/deployment/guides/kubernetes-single.md) -- the Helm chart in `standalone` mode, one pod. | [Split by role, Kubernetes](https://jaque.sh/docs/deployment/guides/kubernetes-distributed.md) -- the chart in `split` mode, one workload per role. |

One node is enough for most installs: [Overview](https://jaque.sh/docs/deployment/overview.md), section
2, is the full argument, and the short version is that the single binary
holds the scheduler, the event log and the dashboard on local disk, with
no operational tax beyond keeping that one process running. An estate
migrating off a single Nagios host almost always fits here, because that
host already proved the capacity is enough.

Splitting by role buys independent scaling and independent failure for
`worker`, `ui`, `sink` and `notifier`, and workers can sit in a network
zone the engine cannot reach directly. It costs a shared log server every
role now depends on, and a second decision -- how many engines -- that
[Cluster and coordination](https://jaque.sh/docs/deployment/cluster-and-coordination.md) covers once more
than one is sharing ownership of the object set.