High Read Latency and aborts in the network for LUNs on AFF System

Applies to

• ONTAP 9.3 & above.
• AFF System
• Environments serving any SAN protocol (iSCSI / FCP / FCoE / NVMEoF)

Issue

• Higher average latency may be observed for a specific LUN/Volume
  • Poor read latency will be reported by user/Application.
  • Potential aborts and timeouts may be seen in host logs.
  • Write latency, however, should not be affected
• The latency breakdown from AIQUM or QOS shows all the delay from Delay-Center:Network.

cluster::> qos statistics volume latency show
Workload            ID    Latency    Network    Cluster       Data       Disk        QoS      NVRAM      Cloud
--------------- ------ ---------- ---------- ---------- ---------- ---------- ---------- ---------- ----------
-total-              -     8.88ms     4.82ms     1.00us  1032.00us     4.02ms        0ms     2.00us        0ms
vm_lun_home1-..   4169   137.73ms   112.91ms     1.00us    22.93ms     1.89ms        0ms        0ms        0ms