ONTAP CLI scripting, schedules & jobs

Build a stable CLI contract

Use -fields for a compact table and -instance for a complete record. Scope commands with supported -vserver and -node parameters; those are command parameters, not global context switches. -show-fulltext true prevents long values being truncated where that parameter is offered. Confirm each command’s parameters with ? or the command reference for your release.

cluster::> volume show -fields vserver,volume,size,used,percent-used
vserver volume size  used   percent-used
svm_nas data01  2TB   1.31TB 65%

cluster::> system node show -node node01 -instance
                              Node: node01
                             Owner: cluster
                            Health: true
                       Eligibility: true

cluster::> event log show -severity ERROR -show-fulltext true
Time                Node   Severity Event
8/27/2026 10:14:02  node01 ERROR    monitor.volume.nearlyFull: ...

Batch operations

The admin shell has no portable shell loop. Obtain a bounded list, then loop on the management host. The example below is client-side shell pseudocode: validate quoting, authentication, host keys, and output on a non-production cluster first.

$ for svm in svm_fin svm_eng svm_archive; do
    ssh ontap-admin "volume show -vserver $svm -fields volume,size,used"
  done
Vserver  Volume   Size   Used
svm_fin  ledger   4TB    2.8TB
...

Prefer the automation guide and REST collection filters for larger estates. REST gives structured JSON and HTTP status codes, avoiding screen-scraping. Pagination, asynchronous job links, TLS validation, and least-privilege RBAC still need explicit handling.

Schedules, snapshots, and asynchronous jobs

Create reusable cron or interval schedules, then attach them through the feature that accepts a schedule. A schedule alone executes no arbitrary CLI command. Here a daily schedule is assigned to a snapshot policy:

cluster::> job schedule cron create -name daily-0130 -minute 30 -hour 1
cluster::> job schedule cron show -name daily-0130
Name        Description
daily-0130  @1:30

cluster::> job schedule interval create -name every-6h -hours 6
cluster::> job schedule interval show -name every-6h
Name      Days Hours Minutes Seconds
every-6h  0    6     0       0

cluster::> volume snapshot policy add-schedule -vserver svm_fin \
  -policy nightly -schedule daily-0130 -count 7
cluster::> volume snapshot policy show -vserver svm_fin -policy nightly
Vserver Policy  Schedules Enabled
svm_fin nightly daily-0130 true

Many mutating commands return a job ID. Capture it and poll job show -id until a terminal state such as Success or Failure. job pause, job resume, and job delete exist, but applicability depends on job state and type; do not delete a record merely to hide a failed operation. There is no general 9.19.1 job start command for arbitrary work.

cluster::> job show -id 1842 -instance
                      Job ID: 1842
                        State: Running
                      Percent: 63%
                  Description: Vol Move svm_fin:ledger
cluster::> job pause -id 1842
1 entry was acted on.

system script start starts recording CLI input and output to a session log; it does not run scripts on nodes. Stop and inspect recordings with system script stop and system script show. EMS event handling can notify or trigger AutoSupport; do not assume an undocumented event script facility exists. Use an external event consumer/webhook platform where supported—check your version.

cluster::> system script start -script-name change-4821
cluster::> volume show -vserver svm_fin -volume ledger
cluster::> system script stop
cluster::> system script show
Script Name  Start Time           Status
change-4821  8/27/2026 11:02:10  complete

Three practical reports

cluster::> volume show -type DP -fields vserver,volume,size,used
vserver volume    size used
svm_dr  ledger_dr 4TB  2.8TB

cluster::> volume show -fields vserver,volume,snapshot-policy
vserver volume snapshot-policy
svm_fin ledger nightly

cluster::> df -h
Filesystem              total used avail capacity Mounted on
/vol/ledger/             4TB  2.8TB 1.2TB 70%      /ledger
cluster::> volume show -fields size,used
vserver volume size used
svm_fin ledger 4TB 2.8TB

Treat displayed tables as human interfaces. If downstream logic depends on types, units, or exact field names, use REST and normalize values explicitly.

Automation safety

• Make operations idempotent: read current state, change only drift, then read back.
• Fail closed on nonzero SSH status, HTTP error, an ONTAP job Failure/Error state, missing rows, or unexpected multiplicity. ONTAP has no universal CLI -error-action; implement error policy in the caller.
• Use explicit SVM, node, volume, and cluster identifiers. Add allowlists and a dry-run mode around destructive commands.
• Expect long jobs and client timeouts. Poll the returned job instead of repeating the mutation.
• Never embed passwords on command lines. Shell metacharacters, spaces, and history expansion can alter or expose them; use supported secret handling.
• Set a known locale and parse machine data. Human output, dates, decimal separators, units, and columns can vary by release or locale.
• Rate-limit parallel work, log request IDs and before/after state, and use dedicated least-privilege accounts.

For orchestration choices, REST and legacy OnCommand tooling, continue with ONTAP automation and keep the CLI cheatsheet nearby.

Official sources

• ONTAP 9.19.1 command reference
• job schedule cron create and interval create
• job show, job pause, job resume, and job delete
• system script start
• ONTAP REST automation documentation