Chapter 3 Statistics and Datasets
Determining the impact of a dynamic statistic
Capacity: Capacity Percent Used
Capacity System Pool Bytes Used
Capacity: System Pool Bytes Used
Capacity System Pool Percent Used
Capacity: System Pool Percent Used
Data Movement NDMP Bytes Statistics
Data Movement: NDMP Bytes Statistics
Data Movement NDMP Operations Statistics
Data Movement: NDMP Operations Statistics
Data Movement Replication Bytes
Data Movement: Replication Bytes
Data Movement Replication Operations
Data Movement: Replication Operations
Data Movement Shadow Migration Bytes
Data Movement: Shadow Migration Bytes
Data Movement Shadow Migration Ops
Data Movement: Shadow Migration Ops
Data Movement Shadow Migration Requests
Data Movement: Shadow Migration Requests
Protocol Fibre Channel Operations
Protocol: Fibre Channel Operations
Protocol: HTTP/WebDAV Requests
Data Movement NDMP Bytes Transferred to/from Disk
Data Movement: NDMP Bytes Transferred to/from Disk
Data Movement NDMP Bytes Transferred to/from Tape
Data Movement: NDMP Bytes Transferred to/from Tape
Data Movement NDMP File System Operations
Data Movement: NDMP File System Operations
Data Movement Replication Latencies
Data Movement: Replication Latencies
Disk ZFS Logical I/O Operations
Disk: ZFS Logical I/O Operations
Memory Kernel Memory Lost to Fragmentation
Memory: Kernel Memory Lost to Fragmentation
This statistic shows SMB Service in Oracle ZFS Storage Appliance Administration Guide operations/sec (SMB IOPS) requested by clients to the appliance. Various useful breakdowns are available: to show the client, filename and latency of the SMB I/O.
See Protocol NFS Operations for an example of a similar statistic with similar breakdowns.
SMB operations/sec can be used as an indication of SMB load, and can be viewed on the Dashboard in Oracle ZFS Storage Appliance Administration Guide .
Use the latency breakdown when investigating SMB performance issues, especially to quantify the magnitude of the issue. This measures the I/O latency component for which the appliance is responsible for, and displays it as a heat map so that the overall latency pattern can be seen, along with outliers. If the SMB latency is high, drill down further on latency to identify the type of operation and filename for the high latency, and, check other statistics for both CPU and Disk load to investigate why the appliance is slow to respond; if latency is low, the appliance is performing quickly, and any performance issues experienced on the client are more likely to be caused by other factors in the environment: such as the network infrastructure, and CPU load on the client itself.
The best way to improve performance is to eliminate unnecessary work, which may be identified through the client and filename breakdowns, and the filename hierarchy view. It's best to enable these breakdowns for short periods only: the by-filename breakdown can be one of the most expensive in terms of storage and execution overhead, and may not be suitable to leave enabled permanently on a busy production server.
|
These breakdowns can be combined to produce powerful statistics. For example:
"Protocol: SMB operations per second of type read broken down by latency" (to examine latency for reads only)
"Protocol: SMB operations per second for file '/export/fs4/10ga' broken down by offset" (to examine file access pattern for a particular file)
"Protocol: SMB operations per second for client 'phobos.sf.fishpong.com' broken down by file name" (to view what files a particular client is accessing)
See Network: Device bytes for a measure of network throughput caused by the SMB activity; Cache: ARC accesses to learn how well an SMB read workload is returning from cache; and Disk: I/O operations for the back-end disk I/O caused.