SMF Type 72 Record (Subtype 4) z/OS 1.13
This table shows the record layout for type 72 SMF records
It's easy to report on SMF 72 data!
Spectrum SMF Writer handles the difficult SMF record parsing for you automatically. You just specify which fields you want to see.
Spectrum SMF Writer also converts the arcane date and time fields and reformats them into an attractive report.
Record length. This field and the next field (total of four bytes) form the RDW (record descriptor word).
Segment descriptor (see record length field).
Bit Meaning When Set
0 New SMF record format
1 Subtypes used
3-6 Version indicators
7 System is running in PR/SM mode
Record type 72(X'48').
Time since midnight, in hundredths of a second, that the record was moved into the SMF buffer.
Date when the record was moved into the SMF buffer, in the form 0cyydddF. See “Standard SMF Record Header” on page 13-1 for a detailed description.
System identification (from the SMFPRMxx SID parameter).
Subsystem identification (‘RMF').
Number of triplets in this record. A triplet is a set of three SMF fields (offset/length/number values) that defines a section of the record. The offset is the offset from the RDW.
Offset to RMF product section from RDW.
Length of RMF product section.
Number of RMF product sections.
Offset to workload manager control section.
Length of workload manager control section.
Number of workload manager control sections.
Offset to service class served data section.
Length of service class served data section.
Number of service class served data sections.
Offset to resource group data section.
Length of resource group data section.
Number of resource group data sections.
Offset to service/report class period data section.
Length of service/report class period data section.
Number of service/report class period data sections.
Offset to response time distribution data section.
Length of response time distribution data section.
Number of response time distribution data sections.
Offset to work/resource manager state section.
Length of work/resource manager state section.
Number of work/resource manager state sections.
Offset to service class period data section.
Length of service class period data section.
Number of service class period data sections.
Offset to swap reason data section.
Length of swap reason data section.
Number of swap reason data sections
RMF version number.
Product name (‘RMF').
Time that the RMF measurement interval started, in the form 0hhmmssF, where hh is the hours, mm is the minutes, ss is the seconds, and F is the sign.
Date when the RMF measurement interval started, in the form 0cyydddF. See “Standard SMF Record Header” on page 13-1 for a detailed description.
Duration of the RMF measurement interval, in the form mmsstttF where mm is the minutes, ss is the seconds, ttt is the milliseconds, and F is the sign. (The end of the measurement interval is the sum of the recorded start time and this field.)
Number of RMF samples.
Bit Meaning When Set
1 Samples have been skipped
2 Record was written by RMF Monitor III
3 Interval was synchronized with SMF 4-15 Reserved.
Sampling cycle length, in the form 000ttttF, where tttt is the milliseconds and F is the sign (taken from CYCLE option). The range of values is 0.050 to 9.999 seconds.
MVS software level (consists of an acronym and the version, release, and modification level - ZVvvrrmm).
Indicates the type of processor complex on which data measurements were taken. Value Meaning 3 9672, zSeries
Bit Meaning When Set
0 The system has expanded storage
1 The processor is enabled for ES connection architecture (ESCA)
2 There is an ES connection director in the configuration
3 System is running in z/Architecture mode
4 IFA processors available.
PR/SM partition number of the partition that wrote this record.
SMF record level change number (X'44' for z/OS V1R6 RMF). This field enables processing of SMF record level changes in an existing release.
Interval expiration time token. This token can be used to identify other than RMF records that belong to the same interval (if interval was synchronized with SMF).
Offset GMT to local time (STCK format).
Offset to reassembly area relative to start of RMF product section.
Length of reassembly area. Area consists of a fixed header and a variable number of information blocks. Length depends on the record type/subtype, but is fixed for a specific type/subtype.
Reassembly area indicator. Value Meaning 0 Record is not broken. 1 Record is broken.Note: This field is used to indicate whether an SMF record is a broken record. Therefore, offset (SMF72RAO) and length (SMF72RAL) are only valid if SMF72RAN = 1. A reassembly area is only present in broken records.
Original interval length as defined in the session or by SMF (in seconds).
SYNC value in seconds.
Projected gathering interval end (STCK format) GMT time.
Sysplex name as defined in parmlib member COUPLExx.
System name for current system as defined in parmlib member IEASYSxx SYSNAME parameter.
Total number of broken records built from the original large record.
Sequence number of this broken record. Every broken record built from the same large record must have a unique sequence number, it is in the range from 1 to SMF72RBR.
Offset to first reassembly information block relative to start of reassembly area header.
Length of reassembly information block.
Number of reassembly information blocks (same value as SMF72TRN in header section).
Total number of sections in the original large record. This field contains information of how many sections of a specific type were contained in the original SMF record. This field is a copy of the number field of the triplet in the original (non broken) record.
Position of the first of one or more consecutive sections described by this block as in the original record. Values in the range of 1 to SMF72RNN are valid for correct processing. A value of 0 will skip processing of this information block. This field provides information where the sections that are part of this broken record were placed in the original record before the split took place. The actual number of consecutive sections contained in this record is available from the actual triplet in the header extension.
Name of active service policy.
Local time/date of policy activation (STCK format).
Service class name.
Service class period number.
Number of users found.
Number of active users found.
Number of active samples (except OUTR).
Number of idle samples.
Number of users delayed for paging at all samples.
Number of users delayed for swapping at all samples.
Number of out and ready users at all samples.
Number of page-ins.
Number of DIV samples.
Total logically swapped samples for the group.
Total swapped samples for the group (except logical).
Total processor using samples for the group.
Total device using samples for the group.
Total processor delay samples for the group.
Total device delay samples for the group.
Total storage delay samples for the group.
Total JES delay samples for the group.
Total HSM delay samples for the group.
Total XCF delay samples for the group.
Total ENQ delay samples for the group.
Total mount delay samples for the group.
Total message delay samples for the group.
Total unknown state samples for the group.
Total valid samples for the group (single state sum of all using, delay, idle, and unknown).
Count of “long” logical swaps for the group.
Count of “long” swaps to expanded storage for the group.
Count of “long” physical swaps for the group.
Number of active frames.
Number of idle frames.
Number of slots used.
Number of DIV frames.
Number of fixed frames.
Number of central frames for all logically swapped users at all samples.
Number of expanded frames for all logically swapped users at all samples.
Number of expanded frames for all swapped users (except logical) at all samples.
Total vector utilization time for the group (microseconds).
Total elapsed time for all transactions that ended in the group. Does not include queued time (microseconds).
Total time spent on JES or APPC queues by all transactions that ended in the group (microseconds).
Number of transactions that ended in the group (microseconds).
Sum of shared page views.
Sum of shared pages in central storage that are valid.
Sum of shared page validations.
Sum of shared page-ins from auxiliary storage.
STOR/OUTR delay samples for swap reason 1: Terminal output wait.
STOR/OUTR delay samples for swap reason 2: Terminal input wait.
STOR/OUTR delay samples for swap reason 3: Long wait.
STOR/OUTR delay samples for swap reason 4: Auxiliary storage shortage.
STOR/OUTR delay samples for swap reason 5: Real storage shortage.
STOR/OUTR delay samples for swap reason 6: Detected long wait.
STOR/OUTR delay samples for swap reason 7: Requested swap.
STOR/OUTR delay samples for swap reason 8: Enqueue exchange swap.
STOR/OUTR delay samples for swap reason 9: Exchange swap.
STOR/OUTR delay samples for swap reason 10: Unilateral swap.
STOR/OUTR delay samples for swap reason 11: Transition swap.
STOR/OUTR delay samples for swap reason 12: Improve central storage usage.
STOR/OUTR delay samples for swap reason 13: Improve system paging rate.
STOR/OUTR delay samples for swap reason 14: Make room for an out too long user.
STOR/OUTR delay samples for swap reason 15: APPC wait.
STOR/OUTR delay samples for swap reason 16: OMVS input wait.
STOR/OUTR delay samples for swap reason 17: OMVS output wait.
STOR/OUTR delay samples for swap reason 18: In-real swap.
The table above is based on the description provided by IBM in its "MVS Systems Management Facilities (SMF)" manual.