IBM TotalStorage DS6000 Series: Performance Monitoring and Tuning

Chapter 14. Copy Services for the DS6000 473

contamwrts Specifies the Number of Contaminating writes for a Concurrent

Copy or XRC volume.

PPRCtrks Specifies the Number of tracks or portion of tracks that were

transferred to the secondary device of a PPRC pair.

NVSspallo Specifies the NVS Space Allocations.

timephread Specifies the Physical Storage Read Response Time in 16 ms

increments.

timephwrite Specifies the Physical Storage Write Response Time in 16 ms

increments.

byteread Specifies the number of Bytes read in 128 KB increments.

bytewrit Specifies the number of Bytes written in 128 KB incremnts.

timeread Specifies the accumulated response time for all read operations.

timewrite Specifies the accumulated response time for all write operation

Example 14-2 Counters displayed with showrank -metrics command

Counter Name Content

byteread Specifies Bytes Read

bytewri t Specifies Bytes Written

Reads Specifies Read Operations

Writes Specifies Write Operations

timeread Specifies Rank Read Response Time

timewrite Specifies Rank Write Response Time

Example 14-3 Counters displayed with showioport -metrics command

Counter Name Content

byteread (FICON/ESCON) ECKD Bytes Read

bytewrit (FICON/ESCON) ECKD Byte Writes

Reads (FICON/ESCON) ECKD Read Operations

Writes (FICON/ESCON) ECKD Write Operations

timeread (FICON/ESCON) ECKD Accumulated Read Time

timewrite (FICON/ESCON) ECKD Accumulated Write Time

bytewrit (PPRC) PPRC Bytes Send

byteread (PPRC) PPRC Bytes Received

Writes (PPRC) PPRC Send Operations

Reads (PPRC) PPRC Received Operations

timewrite (PPRC) PPRC Send Time Accumulated

timeread (PPRC) PPRC Received Time Accumulted

byteread (SCSI) SCSI Bytes Read

bytewrit (SCSI) SCSI Bytes Write

Reads (SCSI) SCSI Read Operations

Writes (SCSI) SCSI Write Operations

timeread (SCSI) SCSI Read Accumulated Time On Channel

timewrite (SCSI) SCSI Write Accumulated Time On Channel

14.11.1 Managing performance with DS CLI

The showrank -metrics output is a good place to begin to monitor your storage subsystem’s

performance. If you can maintain a database with the deltas of these counters, you have a

good basis to spot any significant changes from normal processing. Plan to save these base

counters for the same busy time period each day, and review the delta changes each week.

This discipline will build up an understanding of how your storage subsystem is performing,

and trends or significant changes should provide an early alert for possible I/O processing

bottlenecks.

474 IBM TotalStorage DS6000 Series: Performance Monitoring and Tuning

You can then monitor the logical disks that were defined in each suspect Rank using the

deltas from successive showckdvol -metrics or showfbvol -metrics commands to find which

volume is causing the Rank imbalance.

You should also periodically monitor traffic through your host I/O adapters with the showrank

-metrics command, so that you can identify increases in I/O.