zpool.scrub.run

Start, pause, or cancel a scrub on a ZFS pool.

When action is START, the pool is validated before the scrub begins: the pool must be ONLINE or DEGRADED, must not have an active resilver, and the most recent scrub must be older than threshold days. If any of these checks fail the call returns silently (no error, no alert). The job blocks until the scrub finishes, is paused, or is canceled.

PAUSE and CANCEL skip validation entirely and operate on the pool directly.

At most 10 scrubs may run concurrently across all pools. Attempting to start an 11th raises an error.

On a successful START a ScrubStarted alert is created. If the start fails for a reason other than the threshold or HA checks, a ScrubNotStarted alert is created instead.

Added in version 26.0.0.

This method is a job.

Type: object

Type: array
No Additional Items

Tuple Validation

Parameter 1: data

data

Type: object

Scrub run parameters.

 No Additional Properties

Pool Name

Type: string

Name of the zpool.

Scan Type

Type: enum (of string) Default: "SCRUB"

SCRUB: full data integrity scan. ERRORSCRUB: targeted scan of blocks with known errors.

Must be one of:
  • "SCRUB"
  • "ERRORSCRUB"

Action

Type: enum (of string) Default: "START"

START: begin or resume. PAUSE: pause in-progress scan. CANCEL: stop entirely.

Must be one of:
  • "START"
  • "PAUSE"
  • "CANCEL"

Threshold

Type: integer Default: 35

Days before a scrub is due when the scrub should start.

Result

Type: null


Required roles: POOL_WRITE