Performance data files: QAPMDISK

This database file includes disk file entries and contains one record for each disk resource.

Typically, there is one disk resource per disk unit except for a multipath disk unit that has multiple disk resources associated with it.

Field Name	Description	Attribute
INTNUM	Interval number: The nth sample database interval based on the start time specified in the Create Performance Data (CRTPFRDTA) command.	PD (5,0)
DTETIM	Interval date (`yymmdd`) and time (`hhmmss`): The date and time of the sample interval.	C (12)
INTSEC	Elapsed interval seconds: The number of seconds since the last sample interval.	PD (7,0)
IOPRN	IOP resource name	C (10)
DIOPID	Reserved	C (1)
DSARM	Disk arm number: Specifies the unique identifier of the unit. Each actuator arm on the disk drives available to the machine represents a unit of auxiliary storage. The value of the unit number is assigned by the system when the unit is allocated to an ASP.	C (4)
DSTYPE	Disk drive type such as 9332, 9335, or 6100.	C (4)
DSDRN	Device resource name. Typically, there is one disk (device) resource per disk unit except for a multipath disk unit that has multiple disk resources associated with it (see note 5).	C (10)
DSSCAN	Number of search string commands: This count is always zero, because search string commands are not supported for the 9332, 9335, or 6100.	PD (5,0)
DSBLKR	Number of blocks read: Block is one sector on the disk drive.	PD (11,0)
DSBLKW	Number of blocks written: Block is one sector on the disk drive.	PD (11,0)
DSBLKR	Number of blocks read: The block length is 520 bytes, which includes 8 bytes of system control information.	PD (11,0)
DSBLKW	Number of blocks written: The block length is 520 bytes, which includes 8 bytes of system control information.	PD (11,0)
DSIDLC	Processor idle loop counter (see note 1): The number of times the disk controller passed through the idle loop. This count is increased differently for the 9332 and the 9335. For the 9332, this counter is increased only if the disk controller is totally idle (for example, no I/O operations are active.) For the 9335, even though the disk controller may be idle and the counter gets increased, an I/O operation can be active (for example, seek is being performed.) This field is nonzero for drive types that support a dedicated disk processor and zero for other drive types. DSIDLC and DSIDLT are duplicated across all arms for the same disk controller.	PD (11,0)
DSIDLT	Processor idle loop time (see note 3): The time (in hundredths of microseconds) to make one pass through the idle loop. This field is nonzero for drive types that support a dedicated disk processor and zero for other drive types. The value reported could be a multiple of the actual idle loop time. In that case, the value reported for the processor idle loop count field (DSIDLC) is reduced accordingly so that the calculated processor utilization is correct. DSIDLC and DSIDLT are duplicated across all arms for the same disk controller.	PD (11,0)
DSSK1	Number of seeks > 2/3: The number of times the arm traveled more than 2/3 of the disk on a seek.	PD (11,0)
DSSK2	Number of seeks > 1/3 and < 2/3 (see note 2): The number of times the arm traveled more than 1/3 but less than 2/3 of the disk on a seek.	PD (11,0)
DSSK3	Number of seeks > 1/6 and < 1/3 (see note 2): The number of times the arm traveled more than 1/6 but less than 1/3 of the disk on a seek.	PD (11,0)
DSSK4	Number of seeks > 1/12 and < 1/6 (see note 2): The number of times the arm traveled more than 1/12 but less than 1/6 of the disk on a seek.	PD (11,0)
DSSK5	Number of seeks < 1/12 (see note 2): The number of times the arm traveled from its current position but less than 1/12 of the disk on a seek.	PD (11,0)
DSSK6	Number of zero seeks (see note 2): The number of times the access arm did not physically move on a seek request. The operation may have resulted in a head switch. This field is 0 for disk drive type 6100. The number of zero seeks will be accumulated in DSSK5.	PD (11,0)
DSQUEL	Total queue elements: The number of I/O operations waiting service at sample time. This number includes the I/O operation that is in progress. Divide this by DSSMPL to get the average queue length.	PD (11,0)
DSNBSY	Number of times arm not busy: The number of times there were no outstanding I/O operations active at sample time.	PD (11,0)
DSSMPL	Number of samples taken: The number of samples taken for the DSQUEL and DSNBSY fields.	PD (11,0)
DSCAP	Drive capacity (in bytes): Total number of bytes of auxiliary storage provided on the unit for the storage of objects and internal machine functions when the ASP containing it is not under checksum protection. The unit reserved system space value is subtracted from the unit capacity to calculate this capacity.	PD (15,0)
DSAVL	Drive available space (in bytes): Total number of bytes of auxiliary storage space that is not currently assigned to objects or internal machine functions, and therefore is available on the unit.	PD (15,0)
DSASP	ASP number: Specifies the ASP to which this unit is currently allocated. A value of 1 specifies the system ASP. A value from 2 through 32 specifies a basic ASP. A value from 33 to 255 specifies an independent ASP. A value of 0 indicates that this unit is currently not allocated.	PD (5,0)
DSCSS	Reserved	C (2)
DSPCAP	Reserved	PD (11,0)
DSPAVL	Reserved	PD (11,0)
DMFLAG	' ' means this arm is not locally mirrored. 'A' means this is the designated first arm of a locally mirrored pair. 'B' means this is the designated second arm of a locally mirrored pair.	C (1)
DMSTS	Local mirroring status. 1 = active, 2 = resuming, 3 = suspended	PD (1,0)
DMIRN	Locally mirrored IOP resource name	C (10)
DMDRN	Locally mirrored device resource name	C (10)
DSRDS	Number of read data commands.	PD (11,0)
DSWRTS	Number of write data commands.	PD (11,0)
DSBUFO	Number of buffer overruns: The number of times that data was available to be read into the disk controller buffer from the disk, but the disk controller buffer still contained valid data that was not retrieved by the storage device controller. Consequently, the disk had to take an additional revolution until the buffer was available to accept data. This field is 0 for disk drive type 6100.	PD (11,0)
DSBUFU	Number of buffer underruns: The number of times that the disk controller was ready to transfer data to the disk on a write, but the disk controller buffer was empty. The data was not transferred in time by the disk IOP to the disk controller buffer. The disk was forced to take an extra revolution awaiting the data. This field is 0 for disk drive type 6100.	PD (11,0)
DSMDLN	Model Number: The model number of the disk drive.	C (4)
DSDCRH	Device cache read hits: The number of times that all of the data requested by the read operation was obtained from the device read or write cache.	PD (11,0)
DSDCPH	Device cache partial read hits: The number of times that a portion, but not all, of the data requested by the read operation was obtained by the device read or write cache. A physical operation to the device media was required to obtain the remaining data.	PD (11,0)
DSDCWH	Device cache write hits: The number of times that the data associated with a write operation replaces, or is combined with, existing data in the device write cache, thereby eliminating a write operation.	PD (11,0)
DSDCFW	Device cache fast writers: The number of times that space was available in the device write cache for the data associated with a write operation and a response was returned immediately.	PD (11,0)
DSDROP	Device read operations: The number of read operations issued to the device by the controller. This includes operations generated for redundant system data areas. It does not include operations generated for diagnostics or access to the controller reserved area that occur during this idle time.	PD (11,0)
DSDWOP	Device write operations: The number of write operations issued to the device by the controller. This includes operations generated for redundant system data areas. It does not include operations generated for diagnostics or access to the controller reserved area that occur during this idle time.	PD (11,0)
DSCCRH	Controller cache read hits: The number of times that all of the data requested by the read operation was obtained from the controller read or write cache.	PD (11,0)
DSPCPH	Controller cache partial read hits: The number of times that a portion of the data requested by the read operation was obtained from the controller read and write cache. An operation to the device was required to obtain the remaining data.	PD (11,0)
DSCCWH	Controller cache writes hits: The number of times that the data associated with the write operation replaces or is combined with existing data in the controller write cache. This eliminates a write operation.	PD (11,0)
DSCCFW	Controller cache fast writes: The number of times that space was available in the controller write cache for the data associated with a write operation and a response was returned immediately.	PD (11,0)
DSCOMP	Compressed Unit indicator. '0' if disk data is not compressed and '1' if disk data is compressed.	C (1)
DSPBU	Physical blocks used. For compressed units, this field contains the total number of physical blocks used (written) in the device user data area. For non-compressed units, this field contains 0.	PD (11,0)
DSPBA	Physical blocks allocated. For compressed units, this field contains the total number of physical blocks committed (reserved) in the device user data area for DASD extents. This value includes all of the Physical Blocks Used. For non-compressed units, this field contains 0.	PD (11,0)
DSLBW	Logical blocks written. For compressed units, this field contains the total number of logical blocks written in the device user data area. This value represents the total amount of data written to allocated extents. For non-compressed units, this field contains 0.	PD (11,0)
DSLBA	Logical blocks allocated. For compressed units, this field contains the total number of logical blocks contained in allocated compression groups. This value represents the total sum of all allocated compression groups in the device user data area. For non-compressed units, this field contains 0.	PD (11,0)
DSPBCO	Physical blocks for compression overhead. For compressed units, this field contains the total number of physical blocks that are used for compression directory structures and reserved areas that are unavailable for storing user data. For non-compressed units, this field contains 0.	PD (11,0)
DSFGDR	Foreground directory reads. For compressed units, this field is the number of device read operations that have been performed to read directory structures required to complete host system commands. For non-compressed units, this field contains 0.	PD (11,0)
DSFGDW	Foreground directory writes. For compressed units this is the number of device write operations that have been performed to write directory structures required to complete host system commands. For non-compressed units, this field contains 0.	PD (11,0)
DSBGDR	Background directory reads. For compressed units, this is the number of device read operations that have been performed in the management of compression directory structures, but were not immediately required to complete host system commands. For non-compressed units, this field contains 0.	PD (11,0)
DSBGDW	Background directory writes. For compressed units, this is the number of device write operations. For non-compressed units, this field contains 0.	PD (11,0)
DSFGRE	Foreground read exceptions. For compressed units, this is the number of times an additional device read operation was issued to read data that had been stored in the exception area on a compressed device (this count applies only to multi-page operations). This count reflects only those operations immediately required to complete host system commands.	PD (11,0)
DSFGWE	Foreground write exceptions. For compressed units, this field is the number of times an additional device write operation was issued to write data into the exception area on a compressed device (this count applies only to multi-page operations). This count reflects only those operations immediately required to complete host system commands. For non-compressed units, this field contains 0.	PD (11,0)
DSFGS	Foreground sweeps. For compressed units, a sweep is the process used to store a 1-MB compression group in the correct number of sectors so there are no unused areas in the data region and no used areas in the exception region of the compression group. The number of foreground sweeps is the number of times an entire 1-MB compression group was required to be swept to complete host system commands. The sweep is needed because the data for a host system write operation does not fit into the physical space reserved. The new data does not compress as well as the data that was previously in the space. For non-compressed units, this field contains 0.	PD (11,0)
DSBGS	Background sweeps. For compressed units, a sweep is the process used to store a 1-MB compression group in the correct number of sectors so there are no unused areas in the data region and no used areas in the exception region of the compression group. The number of background sweeps is the number of times an entire 1-MB compression group was swept to maintain the compressed data storage efficiency. This count reflects only those sweeps that were not immediately required to complete host system commands. Background sweeps are intended to increase performance or increase usable capacity of drive. For non-compressed units, this field contains 0.	PD (11,0)
DSCERC	Controller simulated read cache hits: The number of times that all of the data requested by the read operation could have been, but was not, obtained from a controller read cache (not the controller write cache). This field is updated only when Extended Adaptive Cache Simulator is enabled.	PD (11,0)
DSASPN	ASP resource name. Specifies the resource name of the ASP to which this unit is currently allocated. A value of blanks specifies the system ASP or a basic ASP.	C (10)
DSPS	Parity set. The valid value for this field is '1' or '0'. The value of this field is '1' when the disk unit is in a parity set; otherwise, it is '0'.	C (1)
DSHAPS	High availability parity set. The valid value for this field is '1' or '0'. The value of this field is '1' when the disk unit is in a high availability parity set; otherwise, it is '0'.	C (1)
DSMU	Multipath unit. The valid value for this field is '1' or '0'. The value of this field is '1' when the disk resource represents a multipath disk unit (see note 5); otherwise, it is '0'.	C (1)
DSIP	Initial path of multipath unit. The valid value for this field is '1' or '0'. The value of this field is '1' when the disk resource represents the initial path of a multipath disk unit; otherwise it is '0'. The initial path is the first path observed by the system. It can change after restarting the system (IPL). The resource name of the initial path can be used for reporting a multipath disk unit under a single resource name.	C (1)
DSPC	Production copy of remotely mirrored independent ASP. The valid value for this field is '1' or '0'. The value of this field is '1' when the disk unit is in a production copy of a remotely mirrored independent ASP; otherwise, it is '0'.	C (1)
DSMC	Mirror copy of remotely mirrored independent ASP. The valid value for this field is '1' or '0'. The value of this field is '1' when the disk unit is in a mirror copy of a remotely mirrored independent ASP; otherwise, it is '0'.	C (1)
DSRDT	RAID type: type of RAID parity set for this disk unit. The valid value for this field is '1' or '0'. This field only has meaning for disk units in a parity set (DSPS field set to '1'). '0' = RAID 5 parity set, '1' = RAID 6 parity set.	C (1)
DSIOPF	Managed by IOP (See note 6). The valid value for this field is '1' or '0'. The value of this field is '1' when this disk unit is attached to the disk storage adapter which is managed by IOP; otherwise, it is '0'.	C (1)
DSCAT	Disk unit category. This field indicates if this disk unit has some special characteristics, which may require a special interpretation of its performance data. This can also be determined by examining device type and model for this disk unit. X’00’ = no special category applies, X’01’ = this disk unit is located in external storage media.	C (1)

Notes:

9332/9335 inconsistencies:
- 9335 updates the idle count only when the processing unit (A) is not busy. Disk operations such as seek could be in progress. 9332 updates the idle count when there is no activity in any of its processors.
- If there is no movement and no head switch, the 9332 does not count this operation as a zero seek; the 9335 does.
- If an operation causes a head switch (starts a read or write on one track and ends up on another track), the 9332 counts this as a zero seek; the 9335 does not.

9335:  > 2/3                      9332:  >= 2/3                   
                                                                   
       > 1/3  and <= 2/3                 >= 1/3  and <2/3         
                                                                   
       > 1/6  and <= 1/3                 >= 1/6  and <1/3         
                                                                   
       > 1/12 and <= 1/6                 >= 1/12 and <1/6         
                                                                   
                  <= 1/12                < 1/12

The idle loop count and time are used to calculate the storage device controller utilization as follows:
Convert the product of the idle loop count times the idle loop time from hundredths of microseconds to seconds. Subtract this from the interval time, and divide the result by the interval time. For example:

Disk processor utilization = (INTSEC - (DSIDLC * DSIDLT)/10**8)/ INTSEC
The following formulas describe how several of the fields in the previous table can be used to calculate utilization and service time for each arm. For a multipath disk unit, these formulas will give the utilization and service time for each path (resource).
- Arm utilization (DSUTL): The part of the total interval that the arm was being used for I/O operations.
```
DSUTL = Arm Busy = (DSSMPL - DSNBSY)/DSSMPL
```
- Arm accesses per second (DSAS): The number of reads and writes per second for this arm during the interval.
```
DSAS = (DSRDS + DSWRTS)/INTSEC
```
- Service time (DSSRVCT): The average time for an arm I/O operation. This includes disk controller time.
```
DSSRVCT = DSUTL/DSAS
```
  Use the following formula to calculate the service time (DSSTM) for a multipath disk unit, where Xi is the calculated value of X for the i-th path and sum(Xi) is the sum of Xi over all paths:
```
DSSTM = sum(DSSRVCTi * (DSRDSi + DSWRTSi)) / sum(DSRDSi + DSWRTSi)
```
  At low disk utilizations (less than 5%), the service time should be ignored, if this disk unit is managed by IOP (value of DSIOPF field is 1). This is a calculated value based on statistical sampling. When the number of samples is very low, the calculated value may not be accurate.
Performance data is reported for each disk resource that is associated with a multipath disk unit. For a multipath disk unit, the following counters come from the device, which means that their values are duplicated for each disk resource that is reported:
- DSIDLC - Processor idle loop count
- DSIDLT - Processor idle loop time
- DSSK1-6 - Number of seeks
- DSBUFO - Number of buffer overruns
- DSBUFU - Number of buffer underruns
- DSDCRH - Device cache read hits
- DSDCPH - Device cache partial read hits
- DSDCWH - Device cache write hits
- DSDCFW - Device cache fast writers
- DSDROP - Device read operations
- DSDWOP - Device write operations
Other field values that are duplicated include drive capacity (DSCAP) and drive available space (DSAVL).

The arm number (DSARM) and mirror flag (DMFLAG) of a particular multipath disk unit can be used to identify the records associated with that unit.
When data is collected on operating system versions earlier than V5R4, this field is always set to 1, because previous versions cannot determine if the disk unit was IOP-based or not.