Slowness reported for volumes with QOS min configured
Applies to
ONTAP 9
Issue
- Customer reports "NFS not responding" errors on clients for all volumes where QOS_Min is configured.
- Breakdown of latency is seen from Delay Centre : NBlade/Network processing.
qos statistics volume latency show
- High number of NFS Rewinds and Giveups were seen which was driven out of QOS
- Rewind is a mechanism used in Nblade to delay and retry an operation. After retry for a specific number or time, the operations "Giveup".
::*> statistics start -object rw_ctx -counter nfs_rewinds|nfs_giveups|qos_rewinds|qos_giveups|total_rewinds -sample-id <sample-name>
::*> statistics show -sample-id <sample-name>