ONTAP cluster administration basics

A cluster is the management domain; nodes supply compute and storage in HA pairs, while SVMs present data services. Routine administration starts at the cluster shell. Boot and out-of-band environments are separate tools with narrower, higher-risk purposes.

Inventory the management domain

cluster::> cluster show
Node     Health Eligibility
node01   true   true
node02   true   true

cluster::> system node show -fields node,health,eligibility,uptime
node   health eligibility uptime
node01 true   true        61 days 04:12
node02 true   true        61 days 03:58

cluster::> cluster peer show
Peer Cluster Name Availability Authentication
dr-cluster       Available    ok

cluster peer show describes a replication peer, not membership. Use cluster setup interactively on a new node to create or join a cluster; current documentation does not define a routine cluster add-node command. Joining or removing controllers is a planned hardware workflow—check platform and ONTAP-version procedures.

Node lifecycle

Confirm HA state and migrate/stop affected workloads before node power operations. -skip-lif-migration-before-reboot true and inhibit-takeover options remove safeguards; do not copy them into normal runbooks.

cluster::> storage failover show
                              Takeover
Node    Partner Possible State
node01  node02  true     Connected to node02

cluster::> system node reboot -node node01 -reason "CHG-4821 firmware window"
Warning: Are you sure you want to reboot node "node01"? {y|n}: y

cluster::> system node halt -node node01 -reason "CHG-4822 hardware service"
Warning: Are you sure you want to halt node "node01"? {y|n}: y

Never reboot or halt both HA partners casually. Follow the HA takeover/giveback guide and validate cluster time services.

Boot menu and maintenance mode

Reach the boot menu from the physical serial console or SP/BMC console during boot. NetApp’s documented menu includes normal boot, boot without /etc/rc, change password, initialize disks, maintenance mode, restore configuration, install new software, reboot, and configure boot device; exact numbering and availability vary by platform/release. In current documentation, option 5 is maintenance mode—not option 4. Option 4 initializes disks and is destructive.

LOADER> boot_ontap menu
(1) Normal Boot.
(4) Clean configuration and initialize all disks.
(5) Maintenance mode boot.
Selection (1-9)? 5

Maintenance prompt

This is not the cluster CLI. Use it only with a documented procedure or Support guidance. Inventory commands vary; storage show disk is a documented maintenance-mode command. Return with halt, then boot ONTAP from LOADER with boot_ontap.

*> storage show disk
DISK       OWNER    POOL SERIAL NUMBER
0a.00.0    node01   0    S3Z...
*> halt
LOADER> boot_ontap

Service processor and BMC

The SP or BMC is an independent out-of-band controller for remote console, sensors, power control, and recovery when ONTAP networking is unavailable. Hardware determines which name applies. Put it on a restricted management network; it is not a data LIF. RLM is the legacy predecessor.

cluster::> system service-processor show
Node    Type Status  Network Status
node01  BMC  online  autosetup
node02  BMC  online  autosetup

cluster::> system service-processor network modify -node node01 \
  -address-family IPv4 -enable true -dhcp none -ip-address 192.0.2.31 \
  -netmask 255.255.255.0 -gateway 192.0.2.1
cluster::> system service-processor network show -node node01

Use SSH to the SP/BMC address or a serial cable for console access. system node run invokes a node shell command from clustered ONTAP and is not the SP console; legacy node-shell commands should be used only when a current procedure requires them.

Eligibility, votes, and quorum

An eligible cluster node participates in cluster quorum; cluster show -fields eligibility is documented in 9.19.1. Do not depend on a -votes field: it is not in that command’s documented parameters. Quorum protects cluster configuration consistency during partitions. Storage takeover remains an HA-pair function, but loss of cluster quorum can restrict cluster-wide management; plan multi-node outages and interconnect work carefully.

cluster::> cluster show -fields node,health,eligibility
node   health eligibility
node01 true   true
node02 true   true

Configuration backup and upgrade touchpoint

cluster::> system configuration backup create -node node01 -backup-name before_chg4821
Backup before_chg4821.7z created successfully.
cluster::> system configuration backup show -node node01
Node    Backup Name          Size
node01  before_chg4821.7z    48.2MB

Configuration backups protect cluster configuration, not user data. Copy them to an external location using the supported backup settings/download workflow and protect the password. Before an ONTAP image change, follow the full ONTAP upgrade guide; a backup does not replace prechecks, HA validation, or application recovery planning.