High latency from SM Sync on SnapMirror Sync(SM-S) primary volumes
Applies to
- ONTAP 9.9 and earlier
- SnapMirror Synchronous (SM-S)
Issue
- High latency from
SM Sync
on SM-S primary volumes. - The high latency could be observed from several different performance monitoring tools, including:
- CLI
- OCUM
- Harvest/Grafana
- NAbox
- The high latency from
SM Sync
in Workload Latency Breakdown could be much higher(in the range of hundres of seconds) than the total Latency in some cases, which clearly indicates a miscalculation. - High latency seen when Cloud insights WorkLoad Security is turned on for the volume.
- The high latency is usually observed when the IOPS is extremely low.
CLI Example:
cluster1::> qos statistics workload latency show -workload vol1-wid26127
Workload ID Latency Network Cluster Data Disk QoS NVRAM Cloud FlexCache SM Sync
----------- ----- ------ ------ ------ ------ ------ ------ ------ ------ ------ ------
vol1-wid26127 26127 991ms 0.1ms 0ms 0.5ms 1.3ms 0ms 0.2ms 0ms 0ms 35s
vol1-wid26127 26127 918ms 0.2ms 0ms 0.8ms 1.1ms 0ms 0.1ms 0ms 0ms 111s