|
Measures reported by NetCluVolPerfTest
Volumes are provisioned on an aggregate on a cluster node, and the combination of all the volumes constitutes the entire namespace or resource pool for LUNs. Volumes contain file systems that hold user data that is accessible using one or more of the access protocols supported by clustered Data ONTAP, including NFS, CIFS, HTTP, FTP, FC, and iSCSI.
For users to be able to read from/write data into volumes quickly, the I/O requests should be processed rapidly by the volumes. Slowdowns in data retrieval can be attributed to I/O processing bottlenecks. In the event of such slowdowns, administrators need to swiftly isolate the following:
- Which volumes are over-utilized?
- Which volumes are overloaded?
- Which volumes are experiencing serious latencies?
- When were these latencies observed most frequently – while reading or writing?
- What type of operations registered the maximum latency – CIFS, NFS, or iSCSI?
This test provides accurate answers to these questions. With the help of these answers, you can quickly diagnose the root-cause of slowdowns when reading from/writing into a volume.
The measures made by this test are as follows:
| Measurement |
Description |
Measurement Unit |
Interpretation |
| Total_ops |
Indicates the rate at which operations (including read and write) were performed on this volume. |
Ops/Sec |
This measure is a good indicator of how busy the volume is. Comparing the value of this measure across volumes will enable you to quickly detect load-balancing irregularities (if any).
|
| Write_ops |
Indicates the rate at which write operations were performed on this volume. |
Ops/Sec |
|
| Read_ops |
Indicates the rate at which read operations were performed from this volume. |
Ops/Sec |
|
| Avg_latency |
Indicates the average time taken by the WAFL filesystem to process all the operations performed on this volume. |
Milliseconds |
The value of this measure excludes the request processing time and the network communication time of the volume. A high value of this measure is a cause for concern, as it indicates a processing bottleneck.
|
| Read_latency |
Indicates the average time taken by the WAFL filesystem to process the read requests of this volume. |
Milliseconds |
The value of these measures exclude the request processing time and the network communication time of the volume. If the Average latency of a volume is high, then you can compare the value of these measures for that volume to know when the latency occurred – while reading or writing?
|
| write_latency |
Indicates the average time taken by the WAFL filesystem to process the write requests made to this volume. |
Milliseconds |
| Read_data |
Indicates the rate at which data bytes were read from this volume. |
Ops/Sec |
|
| write_data |
Indicates the rate at which data bytes were written to this volume. |
MB/Sec |
|
| Cifs_ops |
Indicates the rate at which the CIFS operations were performed on this volume. |
Ops/Sec |
This measure is inclusive of all the CIFS operations i.e., read, write and other miscellaneous CIFS operations.
By comparing the value of this measure with that of the NFS operations and SAN operations measures for a volume, you can figure out which type of operation imposed the maximum load on that volume. |
| nfs_ops |
Indicates the rate at which the NFS operations were performed on this volume. |
Ops/Sec |
This measure is inclusive of all the NFS operations i.e., read, write and other miscellaneous NFS operations. By comparing the value of this measure with that of the CIFS operations and SAN operations measures for a volume, you can figure out which type of operation imposed the maximum load on that volume.
|
| San_ops |
Indicates the rate at which the SAN operations were performed on this volume. |
Ops/Sec |
This measure is inclusive of all the SAN operations i.e., read, write and other miscellaneous SAN operations. By comparing the value of this measure with that of the CIFS operations and NFS operations measures for a volume, you can figure out which type of operation imposed the maximum load on that volume.
|
| Cifs_latency |
Indicates the average time taken for performing the CIF operations (including read, write and other miscellaneous CIF operations) on this volume. |
Secs |
The value of these measures exclude the request processing time and the network communication time of the volume. Ideally, the value of these measures should be low. If the Avg latency of a volume is very high, then, you can compare the value of these measures for that volume to determine the reason for the latency – is it because of processing bottlenecks experienced by CIFS operations? NFS operations? Or SAN operations?
|
| nfs_latency |
Indicates the average time taken for performing the NFS operations (including read, write and other miscellaneous NFS operations) on this volume. |
Secs |
| San_latency |
Indicates the average time taken for performing the block protocol operations (including read, write and other miscellaneous block protocols operations) on this volume. |
Secs |
|