NS0-185 NetApp Certified Storage Installation Engineer, ONTAP Practice Questions

As-built documentation serves as a critical reference for the customer and support personnel. It captures the exact state of the system at the time of installation, including hardware layout, cabling, network configurations (IPs, VLANs), and software versions. This baseline is essential for effective day-to-day administration, troubleshooting, planning future upgrades, and ensuring consistency.

Each VLAN is a separate layer 2 network, so each needs its own broadcast domain containing that VLAN's port (created with network port vlan create, for example e4a-101) from both nodes; this provides the segmentation. High availability then depends on the protocol: iSCSI (SAN) LIFs do not fail over or migrate—host multipathing (MPIO) moves I/O to another LIF—so iSCSI LIFs are created on both nodes in both the iSCSI A and iSCSI B VLANs. NFS LIFs use LIF failover: the failover group that ONTAP creates for the VLAN 103 broadcast domain contains that VLAN's ports on both nodes, so if a port fails the NFS LIF moves to a surviving port (with the default system-defined policy, the home node and the other node of this two-node cluster). A single broadcast domain for all VLANs mixes layer 2 networks, relying on failover groups for iSCSI LIFs does not work, and placing all LIFs untagged on one interface group removes the segmentation.

The Service Processor (SP) or Baseboard Management Controller (BMC) is a separate, independent computer within the controller. Its primary function is to provide out-of-band management. This allows an administrator to connect to the controller's console, monitor hardware sensors (temperature, voltage), and perform power control operations, even if the main ONTAP operating system is halted, crashed, or unresponsive.

Jumbo frames only help when every hop can carry them. NetApp's guidance is that the data network MTU on the storage system should be 9000 and match the clients and other endpoints, and that 'all components in the network must be able to forward jumbo frames (9000 bytes IP, 9022 bytes including Ethernet)', so data switches are set to at least 9022 bytes (typically 9216). The MTU therefore has to be raised on the client/host NICs, on every switch and router interface in the data path, and in ONTAP on the broadcast domain that contains the data ports (for example network port broadcast-domain modify -broadcast-domain -mtu 9000), which applies it to all ports in that domain. The cluster interconnect ports are not in the client data path; they belong to the Cluster IPspace, whose ports already default to MTU 9000.

Jumbo frames only help when every hop can carry them. NetApp's guidance is that the data network MTU on the storage system should be 9000 and match the clients and other endpoints, and that 'all components in the network must be able to forward jumbo frames (9000 bytes IP, 9022 bytes including Ethernet)', so data switches are set to at least 9022 bytes (typically 9216). The MTU therefore has to be raised on the client/host NICs, on every switch and router interface in the data path, and in ONTAP on the broadcast domain that contains the data ports (for example network port broadcast-domain modify -broadcast-domain -mtu 9000), which applies it to all ports in that domain. The cluster interconnect ports are not in the client data path; they belong to the Cluster IPspace, whose ports already default to MTU 9000.

Jumbo frames only help when every hop can carry them. NetApp's guidance is that the data network MTU on the storage system should be 9000 and match the clients and other endpoints, and that 'all components in the network must be able to forward jumbo frames (9000 bytes IP, 9022 bytes including Ethernet)', so data switches are set to at least 9022 bytes (typically 9216). The MTU therefore has to be raised on the client/host NICs, on every switch and router interface in the data path, and in ONTAP on the broadcast domain that contains the data ports (for example network port broadcast-domain modify -broadcast-domain -mtu 9000), which applies it to all ports in that domain. The cluster interconnect ports are not in the client data path; they belong to the Cluster IPspace, whose ports already default to MTU 9000.

NetApp's switch procedures cable the inter-switch links port-to-port: each ISL port on the first switch connects to the same port number on the second switch (for example, 'Connect ports 1/31 and 1/32 on C1 to the same ports on C2' for Nexus 3232C switches, whose RCF reserves ports 31-32 for the ISLs; BES-53248 switches use ports 0/55-0/56 the same way). The RCF bundles the two links into a port channel, giving two redundant, parallel paths between the switches. Looping two ports of the same switch or landing both links on a single port does not form redundant ISLs, and cross-connecting different port numbers is not the documented cabling.

The correct action is to halt the installation until the site's power infrastructure meets the specifications outlined in the NetApp Hardware Universe. Using underrated power equipment like splitters creates a significant risk of power failure and system instability, which violates installation best practices. The engineer has a responsibility to ensure the environment is suitable to support the hardware reliably.

On FAS8300 and FAS8700 controllers, a two-node switchless cluster interconnect uses the 100 GbE ports: e0c is cabled to e0c and e0d to e0d between the two controllers. The 25 GbE ports e0a and e0b are cabled e0a-to-e0a and e0b-to-e0b as the HA interconnect only, so they are not cluster interconnect connections. e0M is the management (wrench) port, and cluster ports are never connected to a management switch.

On FAS8300 and FAS8700 controllers, a two-node switchless cluster interconnect uses the 100 GbE ports: e0c is cabled to e0c and e0d to e0d between the two controllers. The 25 GbE ports e0a and e0b are cabled e0a-to-e0a and e0b-to-e0b as the HA interconnect only, so they are not cluster interconnect connections. e0M is the management (wrench) port, and cluster ports are never connected to a management switch.

The ANDU process is designed to be nondisruptive, and it will pause if it detects operations that could be impacted. The correct procedure is to identify the source of the lock using ONTAP commands ('vserver cifs session file show' or similar) and resolve the issue at the source (by having the user close the file) rather than forcing a disruptive action or aborting the entire process.

RAID-TEC (triple erasure coding) uses three parity disks per RAID group and protects against up to three simultaneous disk failures in a RAID group, the highest protection ONTAP RAID offers. RAID-TEC is supported on all disk types and platforms, including AFF, even though RAID-DP is the default for all-flash local tiers. RAID-DP protects against two simultaneous failures and RAID4 against one; RAID 0+1 is not an ONTAP aggregate RAID type.

ONTAP clusters can mix platforms. Nodes in an HA pair must be like controller models, but otherwise you can use any supported combination of controllers, and NetApp confirms that AFF and FAS models can be mixed in the same cluster. A 6-node cluster is well within the limits (up to 24 nodes for NAS, 12 for SAN). The cluster base license is a deprecated input and is not required. Confirm the specific platform mix and cluster-switch connectivity in the Hardware Universe, add the new nodes one at a time, and then enable storage failover for the new HA pair.