A type B consistency point (CP) is triggered when the NVRAM buffer is full and needs to be flushed to
disk. A type B CP indicates that the write workload is higher than the disk throughput, and the
system is experiencing back pressure1
A switchless cluster is a cluster configuration that does not use external switches for cluster
interconnect and management network. A switchless cluster has lower bandwidth and redundancy
than a switched cluster, and is limited to two nodes2
The exhibit shows the output of the sysstat -x command, which displays the system performance
statistics in extended mode. The output shows that the system has high CPU utilization, high disk
utilization, high NVRAM utilization, and several type B CPs. These are signs of performance issues
and resource contention3
The best corrective action to address these consistency points is to add an additional shelf of twenty-
four 1.8 TB disks. This will increase the disk capacity and throughput, and reduce the disk utilization
and the frequency of type B CPs4
Creating additional data LIFs will not address the consistency points, because the data LIFs are used
for data access protocols, not for NVRAM flushing5
Converting the 2-node switchless cluster to a 2-node switched cluster will not address the
consistency points, because the cluster interconnect and management network are not related to
the disk performance6
Replacing the twenty-four 1.8 TB disks with twelve 4 TB disks will not address the consistency points,
because the disk throughput will not increase, and the disk utilization will remain high7
Reference:
1: Where can I learn more about Consistency Points? - NetApp Knowledge Base 2: ONTAP 9 - Cluster
and SVM Peering Express Guide - The Open Group 3: ONTAP 9 - Commands: Manual Page Reference -
The Open Group 4: ONTAP 9 - Hardware Universe - The Open Group 5: ONTAP 9 - Network
Management Guide - The Open Group 6: ONTAP 9 - Clustered Data ONTAP Concepts Guide - The
Open Group 7: ONTAP 9 - Logical Storage Management Guide - The Open Group