Poor client performance but ONTAP latency is good - Resolution Guide
Applies to
ONTAP 9
- High-end user latency reported from Internal monitoring software or customer concerns.
- Low throughput/Poor Performance reported by end users
- Data ONTAP and NetApp monitoring tools show low latency when reviewed
Example: vSphere reports high-volume latency or the ESX host reports latency in the log:
2020-09-17T01:31:14.178Z cpu44:2098185)WARNING: ScsiDeviceIO: 1564: Device naa.600a0920002002020202020202020202 performance has deteriorated. I/O latency increased from average value of 744 microseconds to 1515626 microseconds.
Example: qos statistics volume latency show
command shows low latency:
Cluster::> qos statistics volume latency show -volume Example_Vol -vserver VS1 Workload ID Latency Network Cluster Data Disk QoS Max QoS Min NVRAM Cloud FlexCache SM Sync VA --------------- ------ ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- -total- - 2.21ms 101.00us 75.00us 400.00us 1.62ms 0ms 0ms 7.00us 0ms 0ms 0ms 0ms Example_Vol.. 3195 1295.00us 478.00us 0ms 777.00us 26.00us 0ms 0ms 14.00us 0ms 0ms 0ms 0ms
Note: In addition, check the qos statistics volume performance show
command because of offbox latency.
For more information, see article: High latency seen on clients, but not in qos statistics volume latency show