On z/OS, Adabas supports the recording of database activity via SMF records when ADARUN parameter SMF=YES is specified.
SMF recording is supported on z/OS only.
All Adabas SMF records have a common structure, with some sections appearing in all records and others generated according to specific events and parameters specified through ADARUN or operator commands. The ASMFREC macro provides mapping DSECTs for all parts of the SMF record.
This document covers the following topics:
Adabas follows the modern convention for SMF record formats. A single record has:
A standard IBM-type header
A self-defining section that describes a variable number of detail sections
A product ID detail section
User-selected detail sections
Each detail section is described by an eight-byte entry in the self-defining section containing three fields. An entry is also called a triplet.
A 4-byte offset from the beginning of the record to the detail section
A 2-byte count of the number of instances of the detail section
A 2-byte length of each detail section instance. If there are no detail section instances of a given type, the triplet is all zeros.
Field ASNumD in the
product ID section
specifies the number of triplets in the self-defining section.
SMF records are z/OS V-format records with a system-imposed maximum length of 32,756 bytes. Most Adabas SMF records fit within this limit for most reasonable types of ADARUN nucleus specifications. However, detail sections such as File Activity could potentially have several thousand detail section instances.
If the entire set of instances will not fit in the space remaining in
the record, Adabas will include only as many as there are room for and write
the record. The SMF record is reset by clearing the triplets for all detail
sections except the product ID section and then adding as many of the remaining
instances as will fit, repeating until all detail sections are processed. Field
ASSegNo in the product
ID section will start at 1 and be incremented with each additional
record, and field ASSegL will be set to zero for the last (or
only) record for an interval or event.
If any detail section is so large that even one instance would cause the record size limit to be exceeded after resetting the SMF record, that detail type is deactivated.
The header section field ASSTy identifies a record
subtype.
A record of this subtype is generated during nucleus initialization. In addition to the header, self-defining and product ID sections, it contains ADARUN parameter and user sections if these have been selected by SMFDETAIL ADARUN parameter or operator commands.
A record of this subtype is generated during nucleus termination. In addition to the header, self-defining and product ID sections, it contains all detail sections specified by the SMFDETAIL ADARUN parameter or operator commands except for the ADARUN parameter section. Statistics in Adabas detail sections reflect totals for the entire nucleus session.
If interval recording has been specified by the SMFINTERVAL ADARUN parameter or operator command, a record of this subtype is generated at the expiration of each interval. In addition to the header, self-defining and product ID sections, it contains all detail sections specified by the SMFDETAIL ADARUN parameter or operator commands except for the ADARUN parameter section. Statistics in Adabas detail sections reflect activity since the previous interval ended, except where noted. This is also called a delta value.
If the Adabas ADARUN parameter detail section has been specified by the SMFDETAIL ADARUN parameter or operator commands, a record of this subtype is generated whenever an ADARUN parameter value is changed after nucleus initialization. In addition to the header, self-defining, product ID and ADARUN parameter sections, it may also contain a user section if that has been selected by SMFDETAIL ADARUN parameter or operator commands.
The nucleus accumulates usage statistics on the resources it uses to accomplish its tasks. These statistics may be recorded at user or system-defined intervals (see ADARUN parameter SMFINTERVAL) and at termination.
Interval recording (Adabas SMF record subtype 3) provides the usage since the last interval ended for each detail section. Adabas SMF record intervals may be synchronized with one of the system-level intervals specified by PARMLIB member SMFPRMxx entries. This allows straightforward analysis of the usage by allowing direct comparison with other record interval data. For example, you can compare the Adabas interval record with RMF data for the same interval to better understand system performance.
Statistics at termination (Adabas SMF record subtype 2) will have cumulative statistics that reflect activity for the entire nucleus session in each specified detail section.
Every Adabas SMF record contains header, self-defining and product ID sections. You can select additional detail sections through the SMFDETAIL ADARUN parameter or operator commands. Each section is mapped by a DSECT generated by the ASMFREC mapping macro.
The following table summarizes the Adabas SMF record sections:
| Detail Section Description | ASMFREC Macro or ADARUN Parameter Specification | Self-Defining Section Triplet Label Base | ASMFREC DSECT Name Produced by the ASMFREC Macro |
|---|---|---|---|
| Header and self-defining section | --- | --- | ASBase |
| Adabas command activity | CMD | ASTCmd | ASCmd |
| Adabas global cache activity by block type1 | CSHB | ASTChB | ASChB |
| Adabas global cache activity by Adabas file number1 | CSHF | ASTChF | ASChF |
| Adabas global cache activity 1 | CSHG | ASTChG | ASChG |
| Adabas Parallel Services cache activity2 | CSHP | ASTChP | ASChP |
| Adabas file activity | FILE | ASTFile | ASFile |
| Adabas global lock activity1 | LOCK | ASTLok | ASLok |
| Adabas internucleus messaging control block activity | MSGB | ASTMsgB | ASMsgB |
| Adabas internucleus messaging counts | MSGC | ASTMsgC | ASMsgC |
| Adabas internucleus messaging service time histogram | MSGH | ASTMsgH | ASMsgH |
| ADARUN parameter values | PARM | ASTParm | ASParm |
| I/O by DD name | IODD | ASTIODD | ASIODD |
| Product ID3 | ID | ASTPID | ASPID |
| Session Statistics | SESS | ASTSess | ASSess |
| Storage pool | STG | ASTStg | ASStg |
| Thread activity | THRD | ASTThrd | ASThrd |
| zIIP statistics | ZIIP | ASTzIIP | ASzIIP |
| User-defined | USER | ASTUsr | user-defined |
The detail section is available only in cluster environments when either Adabas Cluster Services or Adabas Parallel Services are installed.
The detail section is available only in cluster environments when Adabas Parallel Services is installed.
The product ID section is always included in every SMF record. It may not be specified in the SMFDETAIL ADARUN parameter or in operator commands.
This section describes the different detail record sections:
Note:
The DSECTs provided in the following sections may not be the most
current. To see the most current versions of the DSECTs, generate them using
the ASMFREC
macro.
Adabas Global Cache Activity by Adabas File Number Section: CSHF
Adabas Internucleus Messaging Control Block Activity Section: MSGB
Adabas Internucleus Messaging Service Time Histogram Section: MSGH
IBM has defined a standard format for the initial part of all SMF records in z/OS MVS System Management Facilities (SMF), IBM document SA22-7630. This section begins every Adabas SMF record.
ASBase DSect , Base segment
*
* Standard SMF Header
*
ASRDW DS 0Bl4 Record descriptor word
ASLen DS Bl2 Record length
ASSeg DS Bl2 Segment descriptor
ASFlg DS Bl.8 System indicator flags
ASFStV Equ x'40' Subtypes are valid
ASFV4 Equ x'10' MVS/SP V4 and above
ASFV3 Equ x'08' MVS/SP V3 and above
ASFV2 Equ x'04' MVS/SP V2 and above
ASFVS2 Equ x'02' VS2
ASRTy DS Bl1 Adabas record type
ASTme DS Bl4 Time since midnight when record was +
moved into SMF buffer in 1/100 sec
ASDte DS Pl4 Date when record was moved into SMF +
buffer as 0cyydddF
ASSID DS Cl4 System identifier (SMFPRMxx SID)
ASSSI DS Cl4 Subsystem identifier
ASSty DS Bl2 Subtype
ASStI Equ 1 Adabas initialization
ASStT Equ 2 Adabas termination
ASStS Equ 3 Interval statistics
ASStP Equ 4 Parameter change
ASStA Equ 9 Ad hoc record
*
ASBaseL Equ *-ASBase Length of standard header
The self-defining section follows immediately after the header section. It is part of the header section DSECT.
Each detail section triplet is identified by a base label as shown in the table at the beginning of this section. The base label begins with the prefix specified in the ASMFREC invocation followed by the letter T (for triplet), and then followed by a mnemonic detail section identifier. The base label with suffix O is the offset, with suffix L is the length, and with suffix N is the number of instances.
Here is an example of some triplets.
*
* Self-Defining Section
*
ASSDS DS 0B Self-defining section
* Map of typical section triplet
ASSDSO DS Bl4 Offset to section from start of +
record
ASSDSL DS Bl2 Length of section
ASSDSN DS Bl2 Number of section(s)
Org ASSDS
*
ASTID DS 0Bl8 ID Section (always present)
ASTIDO DS Bl4 Offset to ID section from start +
of record
ASTIDL DS Bl2 Length of ID section
ASTIDN DS Bl2 Number of ID section(s)
*
ASTUser DS 0Bl8 User-Defined Section
ASTUserO DS Bl4 Offset to User-Defined section from+
start of record
ASTUserL DS Bl2 Length of User-Defined section
ASTUserN DS Bl2 Number of User-Defined section(s)
*
ASTParm DS 0Bl8 ADARUN Parameter Section
ASTParmO DS Bl4 Offset to detail section from start+
of record
ASTParmL DS Bl2 Length of each detail section
ASTParmN DS Bl2 Number of detail section(s)
*
. . .
*
ASSDSLn Equ *-ASSDS Length of self-defining section
ASSDSNT Equ ASSDSLn/8 Number of triplets
The product ID section is always present in every Adabas SMF record with one instance. It describes the nucleus generating the SMF record and provides information about the record’s contents.
The 2-byte version code consists of a major version and a minor version. A change, such as adding a new triplet or extending a detail section, will increment the minor version. All existing programs should continue to operate as no existing displacements have changed. A more disruptive change will increment the major version and require existing programs to (at least) be reassembled.
ASPID DSect , Product ID Detail Section +
(always present in SMF record)
ASSMFV DS 0Bl2 SMF record version
ASSMFVM DS Bl1 SMF record major version
ASSMFVN DS Bl1 SMF record minor version
ASSMFVC Equ ASSMFV14 Current version: 1.4
ASSMFV11 Equ x'0101' Version 1.1 - Initial release
ASSMFV12 Equ x'0102' Version 1.2 - Added ALS/ASM
ASSMFV13 Equ x'0103' Version 1.3 - ADARUN parms V8.3
ASSMFV14 Equ x'0104' Version 1.4 - ADARUN parms V8.4
* in coll. seq., new section SESS
ASSegNo DS Bl1 Record segment number
ASSegL DS Bl1 Last segment when = 0
ASNumD DS Bl2 Number of detail type triplets
ASPNm DS Cl8 Product name (ADABAS)
ASVRSC DS Cl8 Product ver/rlse/SM/cum: vvrrsscc
ASSysN DS Cl8 System name
ASSypN DS Cl8 Sysplex name
ASVMN DS Cl8 Virtual machine name
ASJbN DS Cl8 Job name
ASStN DS Cl16 ProcStep/Step name
ASJNm DS Cl8 JES job identifier
ASPgm DS Cl8 Program name
ASGrp DS Cl8 Cluster messaging group name
ASST DS Bl8 Nucleus start time in STCK format
ASIST DS Bl8 Interval start time in STCK format
ASIET DS Bl8 Interval end time in STCK format
ASDBID DS Bl4 Database ID
ASNucX DS Bl2 External nucleus ID
ASNucI DS Bl1 Internal nucleus ID
ASSVC DS Bl1 Adabas SVC number
ASASID DS Bl2 Address space ID
ASASIDI DS Bl4 Reusable address space ID instance
ASComp DS Bl2 Completion code +
x'0ccc' System ABEND code ccc +
x'8ccc' User ABEND code ccc
ASARC DS Bl4 ABEND reason code
*
ASPIDL Equ *-ASPID
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Adabas command activity data is derived from data presented at nucleus shutdown.
There is one instance for each command group: A1/4, BT, CL, ET, E1/4, L1/4, L2/5, L3/6, L9, LF, N1/2, OP, UC, RC, RE, REST, S1/4, S2, S5, S8, S9, YA, YB, YF, YP, YCAL, V1, V2, V3, V4, U0, U1, U2 and U3. There are 34 possible instances but this is subject to change in future releases.
ASCmd DSect , Adabas Command Activity
ASCmdNm DS Cl4 Command name
ASCmdCt DS Bl8 Number of times this command type +
was executed
ASCmdTm DS Bl8 Sum of this command type durations +
in microseconds
*
ASCmdL Equ *-ASCmd
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Adabas file activity data is derived from data presented at nucleus shutdown or in response to a DFILUSE operator command. There is one instance for each file possible in the database as specified by ADADEF MAXFILES up to the highest file number with a non-zero use count. The file number is implied by the sequence number of the instance, starting with zero, which reflects commands such as OP that are not associated with a specific file.
ASFile DSect , Adabas File Activity
ASFileCt DS Bl8 Number of commands executed against +
this file
*
ASFileL Equ *-ASFile
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Global cache statistics are available only for Adabas Cluster Services and Adabas Parallel Services nuclei. They are derived from the ones presented at nucleus shutdown or in response to a DXCACHE operator command. There is one detail section instance for each type of block. Users should examine the block type and not rely on any observed order of the instances. The following block types are reported:
AC: Address Converter
DS: Data Storage
DSST: Data Storage Space Table
FCB: File Control Block
NI: Normal Index
UI: Upper Index
WORK: Protection Data
OTHR: Any other block type
ASChB DSect , Global Cache Activity by Block ASCBCN DS Bl2 Cache Number ASCBRsv1 DS Bl2 Unused ASCBBT DS Cl4 Block type ASCBRT DS Bl8 Reads - Total ASCBRCS DS Bl8 Reads - Completed synchronous ASCBRCA DS Bl8 Reads - Completed asynchronous ASCBRIC DS Bl8 Reads - Data in cache ASCBRNI DS Bl8 Reads - Data not in cache ASCBRFS DS Bl8 Reads - Failed - Structure ASCBRO DS Bl8 Reads - For cast-out ASCBROS DS Bl8 Reads - For cast-out synchronous ASCBROA DS Bl8 Reads - For cast-out asynchronous ASCBWT DS Bl8 Writes - Total ASCBWCS DS Bl8 Writes - Completed synchronous ASCBWCA DS Bl8 Writes - Completed asynchronous ASCBWDR DS Bl8 Writes - Data written ASCBWNR DS Bl8 Writes - Data not written ASCBWSF DS Bl8 Writes - Structure full ASCBVI DS Bl8 Validates issued ASCBVF DS Bl8 Validates failed ASCBBD DS Bl8 Block deletes issued ASCBDR DS Bl8 Deletes reissued due to timeout ASCBUR DS Bl8 Number of times updates redone * ASChBL Equ *-ASChB
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Global cache statistics are available only for Adabas Cluster Services and Adabas Parallel Services nuclei. They are derived from the ones presented at nucleus shutdown or in response to a DXCACHE operator command. There is potentially one instance for each file possible in the database as specified by ADADEF MAXFILES. The size of this detail section precludes the ability to generate it for every possible file, so there is one detail section instance for each file that has non-zero usage. Users should examine the file number and not rely on any observed order of the instances.
ASChF DSect , Global Cache Activity by File ASCFCn DS Bl2 Cache Number ASCFRsv1 DS Bl2 Unused ASCFNum DS Bl4 File number ASCFRT DS Bl8 Reads - Total ASCFRCS DS Bl8 Reads - Completed synchronous ASCFRCA DS Bl8 Reads - Completed a synchronous ASCFRIC DS Bl8 Reads - Data in cache ASCFRNI DS Bl8 Reads - Data not in cache ASCFRFS DS Bl8 Reads - Failed - Structure ASCFRO DS Bl8 Reads - For cast-out ASCFROS DS Bl8 Reads - For cast-out synchronous ASCFROA DS Bl8 Reads - For cast-out asynchronous ASCFWT DS Bl8 Writes - Total ASCFWCS DS Bl8 Writes - Completed synchronous ASCFWCA DS Bl8 Writes - Completed a synchronous ASCFWDR DS Bl8 Writes - Data written ASCFWNR DS Bl8 Writes - Data not written ASCFWSF DS Bl8 Writes - Structure full ASCFVI DS Bl8 Validates issued ASCFVF DS Bl8 Validates failed ASCFBD DS Bl8 Block deletes issued ASCFDR DS Bl8 Deletes reissued due to timeout ASCFUR DS Bl8 Number of times updates redone * ASChFL Equ *-ASChF
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Global cache statistics are available only for Adabas Cluster Services and Adabas Parallel Services nuclei. They are derived from the ones presented at nucleus shutdown or in response to a DXCACHE operator command. This detail section appears with one instance.
ASChG DSect , Global Cache Activity Section ASCGCN DS Bl2 Cache Number ASCGRsv1 DS Bl2 Unused ASCGCOD DS Bl8 Cast-out directory reads issued ASCGCODA DS Bl8 Cast-out directory - async ASCGCODS DS Bl8 Cast-out directory - sync ASCGCOU DS Bl8 Unlock cast-out locks issued ASCGCOUA DS Bl8 Unlock cast-out locks - async ASCGCOUS DS Bl8 Unlock cast-out locks - sync ASCGDR DS Bl8 Directory reads issued ASCGDRA DS Bl8 Directory reads issued - sync ASCGDRS DS Bl8 Directory reads issued - async ASCGPub DS (0*9)Bl8 Publishing requests ASCGSync DS Bl8 Update sync ASCGXEnd DS Bl8 BT/CL/ET transaction end ASCGRedo DS Bl8 Redo threshold ASCGFull DS Bl8 Full buffer pool ASCGAll DS Bl8 All blocks ASCGRABN DS Bl8 Specific RABN ASCGDS DS Bl8 File DS blocks ASCGDSST DS Bl8 DSST blocks ASCGNI DS Bl8 File NI blocks * ASChGL Equ *-ASChG
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Global lock statistics are available only for Adabas Cluster Services and Adabas Parallel Services nuclei. They are derived from the ones presented at nucleus shutdown or in response to a DXLOCK operator command. There is one detail section instance for each lock type. The lock type is implied by the sequence number of the instance, starting with one.
ASLok DSect , Global Lock Section * Lock Types ASLokGC Equ 1 GCB ASLokSE Equ 2 Security ASLokFS Equ 3 FST ASLokUF Equ 4 UFT ASLokSO Equ 5 Save Online ASLokFL Equ 6 Flush ASLokES Equ 7 Global ET Synchronization ASLokRC Equ 8 Recovery ASLokUT Equ 9 UFT-File ASLokIU Equ 10 Index Update ASLokHI Equ 11 Hold ISN ASLokUD Equ 12 Unique DE ASLokET Equ 13 ETID ASLokLT Equ 14 LOB Tracker ASLokCM Equ 15 Command Manager User ASLokDI Equ 16 Data Increment ASLokCP Equ 17 Checkpoint ASLokDT Equ 18 Net-Work DBID Target Assignment ASLokGU Equ 19 Global Update Commmand Sync ASLokPM Equ 20 Parameter ASLokDS Equ 21 DSF ASLokRG Equ 22 RLOG ASLokSP Equ 23 SPATS ASLokCA Equ 24 Cancel ASLokWR Equ 25 TBWK4A/E Table ASLokWU Equ 26 PUTUA/E Table ASLokXI Equ 27 XIDE ASLokRH Equ 28 Replication Handshake ASLokRI Equ 29 Read file/ISN ASLokFA Equ 30 Format AC/AC1 ASLokOC DS Bl8 Obtains - Conditional ASLokOG DS Bl8 Obtains - Granted ASLokOR DS Bl8 Obtains - Rejected ASLokOU DS Bl8 Obtains - Unconditional ASLokOS DS Bl8 Obtains - Synchronous ASLokOA DS Bl8 Obtains - Asynchronous ASLokAC DS Bl8 Alters - Conditional ASLokAG DS Bl8 Alters - Granted ASLokAR DS Bl8 Alters - Rejected ASLokAU DS Bl8 Alters - Unconditional ASLokAD DS Bl8 Alters - Deadlock/Rejected ASLokAS DS Bl8 Alters - Synchronous ASLokAA DS Bl8 Alters - Asynchronous ASLokRL DS Bl8 Releases ASLokRS DS Bl8 Releases - Synchronous ASLokRA DS Bl8 Releases - Asynchronous * ASLokL Equ *-ASLok
This selectable detail section may appear in interval or termination
records (subtypes 2 and 3). Internucleus messaging statistics are available
only for Adabas Cluster Services and Adabas Parallel Services nuclei. They are
derived from the ones presented at nucleus shutdown or in response to a DXMSG
operator command. This detail section appears with one instance. The number of
blocks allocated (ASMsgBBA) and the high water mark
(ASMsgBBH) reflect the entire nucleus session in interval
records.
ASMsgB DSect , Inter-Nucleus Messaging Counts
ASMsgBBA DS Bl8 Message control blocks allocated
ASMsgBBH DS Bl8 Message control blocks used +
(high water mark)
ASMsgBBR DS Bl8 Message control block requests
*
ASMsgBL Equ *-ASMsgB
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Internucleus messaging statistics are available only for Adabas Cluster Services and Adabas Parallel Services nuclei. They are derived from the ones presented at nucleus shutdown or in response to a DXMSG operator command. This detail section appears with one instance. Adabas Parallel Services nuclei report only the count of messages sent.
ASMsgC DSect , Inter-Nucleus Messaging Counts ASMsgCMT DS Cl4 Message type ASMsgCMS DS Bl8 Messages sent ASMsgCMI DS Bl8 Messages incoming (arrived) ASMsgCMA DS Bl8 Messages accepted ASMsgCRS DS Bl8 Replies sent * ASMsgCL Equ *-ASMsgC
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Internucleus messaging statistics are available only for Adabas Cluster Services and Adabas Parallel Services nuclei. They are derived from the ones presented at nucleus shutdown or in response to a DXMSG operator command. This detail section appears with two instances:
The first represents messages subject to the MXMSG timeout parameter.
The second represents certain control messages not subject to MXMSG.
The two instances may be summed for a single representation of all
messages. All message times are in microseconds. The minimum and maximum
durations (ASMsgHMn and ASMsgHMx ) reflect the entire
nucleus session in interval records. Field ASMsgMD2 is an extended
(16-byte) floating point sum of the squares of all message durations. It may be
used to compute a standard deviation.
ASMsgH DSect , Inter-Nucleus Messaging Histogram
ASMsgHXP DS Cl4 Transport service
ASMsgHMM DS Bl4 MXMSG or zero for messages not +
subject to MXMSG
ASMsgHMC DS Bl8 Message count
ASMsgHMD DS Bl8 Sum of all message durations
ASMsgHMS DS Bl16 Sum of squares, all msg durations +
(extended hex floating point)
ASMsgHMn DS Bl4 Minimum duration (us)
ASMsgHMx DS Bl4 Maximum duration (us)
ASMsgHCt Equ 9 Number of histogram buckets
ASMsgHG DS (0*ASMsgHCt)Bl8 Histogram buckets
ASMsgH10 DS Bl8 > 1000 s
ASMsgH09 DS Bl8 > 100 s, <= 1000 s
ASMsgH08 DS Bl8 > 10 s, <= 100 s
ASMsgH07 DS Bl8 > 1 s, <= 10 s
ASMsgH06 DS Bl8 > 100 ms, <= 1 s
ASMsgH05 DS Bl8 > 10 ms, <= 100 ms
ASMsgH04 DS Bl8 > 1 ms, <= 10 ms
ASMsgH03 DS Bl8 > 100 us, <= 1 ms
ASMsgH02 DS Bl8 <= 100 us
*
ASMsgHL Equ *-ASMsgH
This selectable detail section may appear in interval or termination
records (subtypes 2 and 3). Parallel services cache statistics are available
only for Adabas Parallel Services nuclei. They are derived from the ones
presented at nucleus shutdown or in response to a DXCACHE operator command.
This detail section appears with one instance. The directory high water mark
ASCPDHiN and in-use count ASCPDirI reflect the entire
nucleus session in interval records.
ASChP DSect , Parallel Services Cache Activity ASCPCN DS Bl2 Cache Number ASCPRsv1 DS Bl2 Unused ASCPNDir DS Bl8 Number of directory elements ASCPNDiI DS Bl8 Number of directory index elements * Directory Statistics * General ASCPDHiN DS Bl8 High-water mark, this nucleus ASCPDirI DS Bl8 In-use, this nucleus * Read ASCPDRA DS Bl8 Located active ASCPDRF DS Bl8 Obtained from free pool ASCPDRC DS (0*4)Bl8 Reclaim criteria categories ASCPDNN DS Bl8 First choice criteria ASCPDND DS Bl8 Second choice criteria ASCPDIN DS Bl8 Third choice criteria ASCPDID DS Bl8 Fourth choice criteria ASCPDCF DS Bl8 Unable to obtain (cache full) ASCPDRT DS Bl8 Tested for reclaim * Write ASCPDWF DS Bl8 Obtained from free pool * Space Management Statistics * Request Statistics ASCPSRP DS Bl8 Sufficient preallocated space ASCPSRF DS Bl8 Free space allocated ASCPSRN DS Bl8 Reclaim space, first choice ASCPSRI DS Bl8 Reclaim space, second choice ASCPSRU DS Bl8 Space unavailable (cache full) ASCPSSP DS Bl8 Searched part of space chain ASCPSSF DS Bl8 Searched entire space chain ASCPSST DS Bl8 Number of space seqs tested * Element Reclaim Statistics ASCPSEN DS Bl8 First choice criteria ASCPSEI DS Bl8 Second choice criteria * Latch management statistics * Cache Space Chain ASCPSPGE DS Bl8 Get Exclusive ASCPSPWF DS Bl8 WaitFor Exclusive ASCPSPRE DS Bl8 Release Exclusive * Cache Directory Index ASCPDIGE DS Bl8 Get Exclusive ASCPDIGS DS Bl8 Get Shared ASCPDIUE DS Bl8 Upgrade Exclusive ASCPDIWE DS Bl8 WaitFor Exclusive ASCPDIWS DS Bl8 WaitFor Shared ASCPDIWU DS Bl8 WaitFor Upgrade ASCPDIRE DS Bl8 Release Exclusive ASCPDIRS DS Bl8 Release Shared * Cache Directory ASCPDRGE DS Bl8 Get Exclusive ASCPDRGS DS Bl8 Get Shared ASCPDRUE DS Bl8 Upgrade Exclusive ASCPDRWE DS Bl8 WaitFor Exclusive ASCPDRWS DS Bl8 WaitFor Shared ASCPDRRE DS Bl8 Release Exclusive ASCPDRRS DS Bl8 Release Shared * Cast-Out Class ASCPCOGE DS Bl8 Get Exclusive ASCPCOGS DS Bl8 Get Shared ASCPCOWE DS Bl8 WaitFor Exclusive ASCPCOWS DS Bl8 WaitFor Shared ASCPCORE DS Bl8 Release Exclusive ASCPCORS DS Bl8 Release Shared * ASChPL Equ *-ASChP
This selectable detail section may appear in initialization records or whenever an ADARUN parameter is changed while the nucleus is running (subtypes 1 and 4 in the header section). It will not be generated for interval or termination records (subtypes 2 and 3). This section has a fixed-length portion containing most parameters, followed by variable-length areas for parameters capable of multiple values or lists of values.
Where possible, the individual field names are formed by prefixing the shortest allowable form of the parameter with ASP. In general, the SMF record will report character parameters in EBCDIC and numeric parameters in binary.
Parameters with limited enumerated values (YES or NO, for example) are reported in 1-byte fields if the possible values are unambiguous. Otherwise, the field length is that used by the nucleus, usually 4 bytes.
Here are some sample entries:
ASParm DSect ,
ASPAO DS Cl1 AOslog
ASPARE DS Bl4 ARExclude
Offset to file table
ASPARMN DS Cl16 ARMname
ASPASS DS Cl1 ASSocache
ASPASY DS Cl1 ASYtvs
. . .
ASPVI DS Cl1 Vista
ASPV64B DS Cl1 V64Bit
ASPWO DS Cl1 Workcache
ASParmV DS 0B Begin variable part
ASParmL Equ *-ASParm Length of ID section
ASParm DSect , ADARUN Parameters * * Adabas Nucleus Parameters * ASPAO DS Cl1 AOslog ASPARE DS Bl4 ARExclude Offset to file table ASPARMN DS Cl16 ARMname ASPASS DS Cl1 ASSocache ASPASY DS Cl1 ASYtvs . . . ASPVI DS Cl1 VIsta
The ADARUN AREXCLUDE parameter is a variable length list of values. The base parameter entry will be an offset from the beginning of the detail section to the table of values. The table is a 4-byte inclusive field followed by 4-byte file numbers. A separate DSECT maps the AREXCLUDE file exclusion table:
ASPAFE DSect , ARM File Exclusion Table ASPAFEN DS Bl4 Inclusive length of table ASPAFEF DS 0Bl4 First file number entry
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). I/O by DD data is derived from data presented at nucleus shutdown. There is one instance for each DD statement administered by the nucleus. You should examine the DD name and not rely on any observed order of the instances. An Adabas nucleus may open and close the same DD name multiple times. Multiple uses of a DD name are summed.
You might see this many DD statement entries in a single nucleus:
| Statement Type | Number of DD Statements |
|---|---|
| ADC | 248 |
| ADP | 248 |
| ADW | 62 |
| ASSO | 99 |
| CLOG | 8 |
| DATA | 99 |
| ECS | 1 |
| PLOG | 8 |
| RLOG | 1 |
| WORK | 2 |
DD names beginning with "AD" represent files that were dynamically allocated, such as during Cluster or Parallel Services recovery, after another nucleus failed.
ADC and ADP represent CLOGs and PLOGs from other nuclei. There may be one set of 8 for each of the 31 possible other nuclei up to a maximum of 248.
Similarly, ADW represents a WORK data set from another nucleus. There may be two for each of the 31 possible other nuclei up to a maximum of 62.
ASIODD DSect , I/O Activity by DD ASIODDNm DS Cl8 DD Name ASIODDRd DS Bl8 Reads ASIODDWt DS Bl8 Writes * ASIODDL Equ *-ASIODD
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Session statistics data is similar to the data presented at nucleus shutdown or at the DSTAT operator command.
There is one single instance per record. The section has a fixed-length.
Here is a shortened sample of the SESS section:
ASSess DSect , Nucleus session statistics * ASSDURA DS BL8 Nucleus Session Duration ASSWAIT DS BL8 Nucleus Wait Time ASSCPU DS BL8 Nucleus CPU Time ASSASR DS BL8 ASSO Reads ASSASW DS BL8 ASSO Writes ... ASSBFFE DS BL4 Buffer efficiency .x precision ASSessL Equ *-ASSess Length of the session section
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Storage pool statistics are derived from statistics presented at nucleus shutdown or in response to a DRES operator command. There is one instance for each storage pool with a non-zero size. Be sure to examine the pool name and not rely on any observed order of the instances.
Storage pool statistics are reported two ways: in bytes and also in units such as a user might specify as an ADARUN parameter, for example, NC. When the units are bytes, the two sets of statistics are the same.
Normally an interval record would show the change from the previous interval, but that isn’t meaningful for storage pools. Thus the interval and termination record subtypes all reflect total usage for the nucleus session.
ASStg DSect , Storage Pool Usage ASStgNm DS Cl4 Storage pool name ASStgBSz DS Bl8 Size in bytes ASStgBHW DS Bl8 High water mark in bytes ASStgUSz DS Bl8 Size in units from ADARUN parameter ASStgUHW DS Bl8 High water mark in ADARUN units * ASStgL Equ *-ASStg Length of Storage Pool section
These are the possible storage pools:
| Storage Pools | ||
|---|---|---|
| Pool Name | ADARUN Parameter | Description |
| AB | NAB | Attached buffers |
| CQ | NC | Command queue |
| DUQ | LDEUQP | Unique descriptor |
| FI | LFP | Internal format buffers |
| HQ | NH | Hold queue |
| PLIO | NPLOGBUFFERS | PLOG I/O buffers |
| REDO | LRDP | Deferred publishing |
| RPL | LRPL | Replication pool |
| SC | LCP | Security information |
| TBI | LI | ISN table |
| TBS | NQ | Sequential command table |
| UQ | NU | User queue element |
| UQF | NU | User queue file elements |
| WKIO | NWORK1BUFFERS | Work I/O buffers |
| WORK | LWP | Work |
| WKP1 | LP | Work part 1 |
| WK1B | CLUWORK1CACHE | Work part 1b |
| WKP2 | LWKP2 | Work part 2 |
| WKP3 | -- | Work part 3 |
| XID | NU | Transaction ID |
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). Thread activity data is derived from data presented at nucleus shutdown or in response to a DTH operator command. The ADARUN parameter NTHREAD defines the number of user threads for the nucleus session. There is one instance for each defined user thread. The thread number is implied by the sequence number of the instance.
ASThrd DSect , Thread Activity
ASThrdCt DS Bl8 Number of commands executed in this +
thread
*
ASThrdL Equ *-ASThrd
This selectable detail section may appear in interval or termination records (subtypes 2 and 3). zIIP activity data is derived from data presented at nucleus shutdown or in response to a DZSTAT operator command. There is one instance for each zIIP-enabled nucleus product.
ASzIIP DSect , Nucleus session statistics * ASzProd DS Cl4 Sub-product name ASzPNuc Equ c'ANUC' Adabas nucleus main task ASzPRevL Equ c'REVL' Review local ASzRsv1 DS Bl4 Unused * ASz2SRB DS Bl8 Count of switches to SRB mode ASz2TCB DS Bl8 Count of switches to TCB mode ASzPAUT DS Bl8 Count of pause TCB requests ASzRLST DS Bl8 Count of release TCB requests ASzPAUS DS Bl8 Count of pause SRB requests ASzRLSS DS Bl8 Count of release SRB requests ASzQPRQ DS Bl8 Count of queue parallel requests ASzXRQE DS Bl8 Count of no free RQE for requests * ASzNCtr Equ 64 Number of event counters ASzCtr DS (ASzNCtr)Bl8 Event counters * ASzCPUT DS Bl8 CPU time (STCK format) ASzZTim DS Bl8 zIIP time (STCK format) ASzSCPU DS Bl8 SRB-on-CPU time (STCK format) ASzZQCP DS Bl8 zIIP-qualified time (STCK format) * ASzIWait DS Bl8 ADAIOR general WAITs ASzIRlse DS Bl8 ADAIOR releases * ASzIIPL Equ *-ASzIIP Length of the zIIP section
Use the ASMFREC macro to generate the latest SMF record DSECTs. The ASMFREC macro will always generate the header and self-defining section DSECT. Detail section DSECTs will be generated as specified. The header and self-defining sections are mapped by a single DSECT. Each detail section is mapped by its own DSECT. The syntax of the ASMFREC macro is:
label ASMFREC Prefix={AS | prefix},
Detail={All | (type [,type]...)},
Title = {'Adabas SMF Records' | 'string'}
Specify a character string to be used as the initial characters for all DSECT and field names. The default is Prefix=AS.
Identify which detail section DSECTs are to be included in the expansion. "All" is the default and will include all detail sections. Alternatively, a comma-delimited list of types (enclosed in parentheses) can be specified; only the types specified will be included. The valid types are shown in the following table. A null value (Detail=) will inhibit all detail section DSECTs.
| ASMFREC Macro Specification | Detail Section Description | ASMFREC DSECT Name Produced by the ASMFREC Macro1 |
|---|---|---|
| CMD | Adabas command activity | xxCmd |
| CSHB | Adabas global cache activity by block type | xxChB |
| CSHF | Adabas global cache activity by Adabas file number | xxChF |
| CSHG | Adabas global cache activity | xxChG |
| CSHP | Adabas Parallel Services cache activity | xxChP |
| FILE | Adabas file activity | xxFile |
| LOCK | Adabas global lock activity | xxLok |
| MSGB | Adabas internucleus messaging control block activity | xxMsgB |
| MSGC | Adabas internucleus messaging counts | xxMsgC |
| MSGH | Adabas internucleus messaging service time histogram | xxMsgH |
| PARM | ADARUN parameter values | xxParm |
| IODD | I/O by DD name | xxIODD |
| SESS | Nucleus session statistics | xxSess |
| STG | Storage pool | xxStg |
| THRD | Thread activity | xxThrd |
| ZIIP | zIIP statistics | xxzIIP |
| USER | User-defined | user-defined |
Where xx is the prefix specified in the ASMFREC macro.
If the Title default "Adabas SMF Record " or another quoted string is specified, an assembler Title statement is generated before the header section DSECT. A null value (Title=) for this operand will inhibit a title in the DSECT.
You can provide a user exit if you want to add a detail section to the Adabas SMF record. The user exit is a separate load module whose name must be provided in the ADARUN UEXSMF parameter. For complete information about the user exit, read SMF User Exit, in the Adabas User, Hyperdescriptor, Collation Descriptor, and SMF Exits Manual.
An Adabas nucleus can register with z/OS to have CPU usage statistics included in IBM type 89 SMF records. These records are described in z/OS MVS System Management Facilities (SMF), IBM document SA22-7630.
To activate type 89 recording for Adabas, specify ADARUN parameters SMF=YES and SMF89=YES. During initialization Adabas will register the nucleus address space with z/OS SMF and have its CPU statistics included in subtype 1 of the type 89 records. The address space is deregistered at nucleus termination. Each Adabas nucleus appears as a separate type 89 entry.
The type 89 entries include CPU usage and a number of descriptive registration parameters. Adabas nuclei use these descriptive fields in type 89 entries as follows:
| SMF Type 89 Descriptive Fields | ||||
|---|---|---|---|---|
| Name | Length | Format | Description | Value |
| SMF89UPO | 16 | EBCDIC | Product owner or vendor name | SOFTWARE AG |
| SMF89UPN | 16 | EBCDIC | Product name | ADABAS |
| SMF89UPV | 8 | EBCDIC | Product version | The eight-byte product version has two-byte numeric values for the Adabas version, release, SM level, and cumulative level. |
| SMF89UPQ | 8 | Binary | Product qualifier |
The product qualifier is a seven-byte string that may be used to distinguish among several nucleus instances. It contains a series of binary fields:
SVC (1 byte) Use both the SVC and DBID to identify instances of Adabas Cluster or Parallel Service nuclei for the same database on any one system. |
| SMF89UPI | 8 | EBCDIC | Product ID |
The product ID is a string of up to eight single characters to show what add-on products are being used. The characters may appear in any order:
C (Adabas Cluster
Services) The following applies:
C, D, P, and R are mutually exclusive. |