NetApp ONTAP: Troubleshoot Ethernet Ports with ifstat

When a NetApp storage node starts dropping packets or a client reports poor NFS throughput, the first place to look is the network port counters — and on ONTAP the primary tool is ifstat. Available both in Data ONTAP 7-mode and from the node-shell of ONTAP 9 clusters, it exposes the device-level statistics that separate a cable problem from a driver problem from a switch problem. This guide explains how to read and clear ifstat counters, what each error counter means, and how to judge whether a counter value is actually a problem.

The ifstat Command

::*> system node run -node cm2520-ks6-01 -command ifstat e0a

Options: ifstat -a shows all interfaces (including virtual host and loopback), ifstat -z e0a clears counters for one port, ifstat -a -z clears all. Counters are cumulative since boot or the last reset — a giveback does not zero them.

Reading the Output

Key receive-side counters to watch:

• CRC errors — physical layer corruption; typically a bad cable, SFP, or negotiated speed mismatch.
• Runt frames / Fragment — frames shorter than minimum; common with duplex mismatches.
• Long frames / Jabber — oversized frames; can indicate MTU or driver issues.
• No buffer — the NIC ran out of buffers under load.
• Bad TCP/UDP checksum — often points to offload configuration or an upstream corruption source.

Transmit side: Late collisions indicate a duplex mismatch or too-long cable segment; Total errors on TX should normally stay at zero. The LINK INFO section shows speed, duplex and flow control — a port stuck at a lower speed or half-duplex explains most throughput complaints.

Interpreting Counters Correctly

A large error count is only meaningful relative to time: if the counters have been running for months without a reset, even thousands of errors may be noise. Check the rate — clear the counters with ifstat -z, reproduce the issue, and compare the delta. Also check Current state: down — a down port means the problem is physical or on the peer side.

Related Network Diagnostics

Combine ifstat with the ONTAP network diagnostic commands: network ping, network traceroute, network tcpdump start, and network test -path for intercluster path latency. See our guides on monitoring ONTAP network port reachability and responding to degraded ONTAP system health for the full health workflow.

原文链接：https://kb.netapp.com/on-prem/ontap/Ontap_OS/OS-KBs/How_to_troubleshoot_Ethernet_Network_Ports_in_ONTAP_using_ifstat