This document describes the initialization parameters needed for an Adabas Parallel Services environment. It covers the following topics:
ADARUN parameters are used to customize the Adabas environment. For information about ADARUN parameters and the format of ADARUN control statements, read your Adabas operations documentation.
For Adabas Parallel Services environments, two types of initialization
parameters must be set based on the ADARUN parameter
PROGRAM.
| Parameter | Specify... | Possible Values | Default |
|---|---|---|---|
PROGRAM |
the ADARUN program | ADANUC | ADACOM | none |
The PROGRAM parameter specifies the name of the
program being run, and must always be specified. There is no default.
To run an Adabas nucleus, the name must be "ADANUC".
To set up an ADACOM task, the name must be "ADACOM" (no other ADARUN parameters are recognized for the ADACOM program run).
For example, the following example starts an Adabas Parallel Services nucleus:
ADARUN PROG=ADANUC,CLUSTER=LOCAL,NUCID=3, ...
An unlimited number of Adabas Parallel Services nucleus clusters, each with up to 31 members sharing a common database, can be defined for an operating system image.
The ADACOM initialization task (ADARUN PROGRAM=ADACOM) must
be run in order to set up the environment, and it must be maintained in order
to monitor and control the nuclei of one or more Adabas Parallel Services
clusters.
ADACOM initialization parameters specify the Router ID / DBID combinations (sets) that the ADACOM is to manage.
The Router ID identifies the SVC number on OS/390, z/OS, or VSE, or the IDTNAME on BS2000/OSD. The Router ID value must be the same within a cluster; however, the same Router ID may be used for different clusters.
The DBID identifies the external physical database shared by a particular cluster of nuclei and known to the application.
Other ADACOM parameters are discussed in the following sections.
Although a single ADACOM job can run all Router ID / DBID sets in an Adabas Parallel Services environment, it is possible to run multiple ADACOM tasks simultaneously with the same, mixed, or completely different Router ID / DBID sets. An ADACOM subtask is attached to each Router ID / DBID set for each ADACOM task in which it occurs.
ADACOM prints global messages that apply to all Router ID / DBID sets to
the data set or file with DD name/link name COMPRINT. Furthermore, ADACOM
prints messages for each individual Router ID / DBID set to an output data set
or file with the DD name/link name
Pssddddd, where
ss is the last two digits of the SVC number and
ddddd is the DBID. (On BS2000, the
ss number is derived for each separate IDTNAME as
the first available, counting backward from 255.) On OS/390 and z/OS, ADACOM
automatically allocates this data set in the spool with SYSOUT=*, if it is not
explicitly specified.
For a sample job to run ADACOM, read the OS/390 or z/OS section Create a Startup Procedure for ADACOM.
| Parameter | Specify... | Minimum | Maximum | Default |
|---|---|---|---|---|
DBID
|
the database to be used. | 1 | 65000 | none |
The DBID parameter designates the ID of an
Adabas Parallel Services cluster's external physical database; that is, the
database number that the user programs address to send commands to the single
physical database of an Adabas Parallel Services nucleus cluster.
This number must be unique among all Adabas database IDs, NUCIDs, Natural buffer pool IDs, etc.
| Parameter | Specify... | Possible Values | Default |
|---|---|---|---|
FORCE
|
whether ADACOM forces a restart after an abnormal termination. | YES | NO | NO |
If Adabas Parallel Services believes there are still active nuclei,
ADACOMs, or users on an image after a termination, a NU
parameter value that is different from the NU value already in effect will not
be recognized on restart. If you are certain that Adabas Parallel Services is
wrong in its belief, you can use FORCE=YES to force a
clean restart.
Note, however, that if a cluster nucleus or an ADACOM with the same SVC
number or IDTNAME is active on the image where FORCE=YES
is used, it will crash.
Notes:
FORCE=YES works only when the NU parameter
value is being changed and has the effect of resetting the environment.
FORCE is
neither seen nor used by the SVCCLU. The ADACOM
parameter FORCE may only be used to change the
NU that is valid for the cluster.
FORCE parameter,
remember to remove it from ADACOM after you have reset the environment.
| Parameter | Specify... | Possible Values | Default |
|---|---|---|---|
IDTNAME
|
the IDT common memory pool name to be used for the Adabas Parallel Services cluster session. | ADAiiiii | ADABAS5B |
The IDTNAME defines the name of the common
memory pool used to find communication details for this nucleus. This common
memory pool will be defined as Global.
Value is "ADAiiiii", maximum of eight characters.
For example, IDTNAME=ADAPROD1
| Parameter | Specify... | Minimum | Maximum | Default |
|---|---|---|---|---|
NU
|
the number of users that can be active in parallel on cluster nuclei in the image. | 20 | 16 777 215 | 200 |
In the extended CSA (ECSA), the SVCCLU maintains
a user table with entries (UTEs) containing information about every active user
in the cluster nuclei on the operating system image. A UTE is assigned when a
user issues an OP command or (if the user did not issue an OP command) at the
first Adabas command. The UTE is released when the user issues a CL command or
when a timeout occurs.
The ADACOM NU parameter specifies the number of
concurrent users (UTEs) allowed for all the nuclei of a cluster. The first
ADACOM started governs the value for NU: different
values set for subsequent nuclei or ADACOMs are ignored.
To manually change the NU value, you must stop
all cluster nuclei and ADACOM tasks in the image, modify the
NU value for all the nucleus and ADACOM jobs, and then
restart.
When the NU parameter is set to zero, any space
allocated for the user table is freed, the Adabas Parallel Services control
block is freed, and the ADACOM task terminates. However, if ADACOM believes
that there are active nuclei, ADACOMs, or users, and
FORCE=YES is not used, no action is taken.
For example, the following parameter would run the Adabas Parallel Services cluster nuclei with 500 elements in the user table:
NU=500
| Parameter | Specify... | Possible Values | Default |
|---|---|---|---|
SVC
|
the supervisor call number to be used for the Adabas Parallel Services nucleus cluster session. |
OS/390 or z/OS: 200-255 VSE: see text |
OS/390 or z/OS: 249 VSE: 45 |
The SVC is used to perform various Adabas
internal functions; the number is used to communicate between the users and the
database.
The SVC number is specified as an integer and must correspond to the number used for the version 7.4 Adabas SVC (ADASVC).
For OS/390, valid SVC values are 200-255.
For VSE/ESA, 45 is the recommended value but any free SVC value can be used. Read the Adabas installation documentation for information about finding free VSE Adabas SVC values.
For example, the following parameter will execute an Adabas Parallel Services nucleus cluster session on an OS/390 image using ADASVC 202.
SVC=202
Software AG recommends that you use the default settings (or your existing values) of the Adabas ADARUN parameters for each Adabas nucleus in an Adabas Parallel Services cluster, and then tune the values after analyzing the performance of the node or cluster. Read Performance and Tuning for information about expected differences.
Session statistics can be used to determine the best settings for each parameter. The statistics are printed automatically at the end of a session, but can also be displayed using nucleus or ADACOM operator commands during the session.
This section describes the ADARUN parameters used to invoke a cluster
nucleus (ADARUN PROGRAM=ADANUC).
ADARUN parameters that must be the same for all nuclei in the cluster are called "global".
Some global parameters are set at nucleus startup and cannot be changed during the ensuing session; other global parameters can be changed during a session.
After the first nucleus in an Adabas Parallel Services cluster starts, message services are used to communicate the ADARUN parameter settings of the first nucleus to all subsequent cluster nuclei. Each following nucleus receives this information during initialization and determines whether its global nonchangeable parameters are equal to those of the first nucleus.
If they are not equal, the nucleus fails with a parameter error. The nonequal global changeable parameters are reset to the value retrieved from the message services and a corresponding message is printed.
On a running system, a cluster nucleus may want to modify one or more of the "changeable" global parameters. This nucleus acquires a "parameter change lock", makes the changes in its local parameter area, and communicates the changes to the other cluster nuclei using message services.
All other nuclei in the cluster receive the messages containing the global parameters that have changed, change the parameters in their local parameter area, and send the "acknowledge" message.
A cluster nucleus requires
global parameters. Adabas Parallel Services enforces the
same value for all nuclei in a cluster. Some of these parameters are modifiable
(GM) during a session using an operator command or AOS
(NISNHQ, NONDES, and
AOSLOG are only modifiable using AOS); others are fixed
(GF) and cannot be modified.
local parameters, which can be different for each nucleus. Some of these parameters are modifiable (LM) using an operator command or AOS; others are fixed (LF) and cannot be modified.
A few Adabas ADARUN parameters are not available or applicable to a cluster nucleus (No).
Note:
All Adabas Caching Facility ADARUN parameters are supported as
local, modifiable parameters.
| Parameter | Usage | No* | LF | LM | GF | GM |
|---|---|---|---|---|---|---|
AOSLOG
|
Log to DDPRINT commands issued by AOS or ADADBS OPERCOM that modify the active nucleus | Y | ||||
AREXCLUDE
|
Exclude file(s) from autorestart | Y | ||||
ARMNAME
|
Name used to activate ARM | Y | ||||
ASSOCACHE
|
Controller caching control for the Associator component | Y | ||||
ASYTVS
|
Asynchronous buffer flush based on vol-ser | Y | ||||
CDXnn
|
Collation descriptor user exit(s) | Y | ||||
CLOGDEV
|
Multiple command log device type | Y | ||||
CLOGLAYOUT
|
Define command log format | Y | ||||
CLOGMRG
|
Automatic command log merge control in a cluster environment | Y | ||||
CLOGSIZE
|
Multiple command log size (blocks) | Y | ||||
CLUCACHENAME
|
Cluster cache structure name (Cluster Services only) | Y | ||||
CLUCACHESIZE
|
Cluster cache area size (Parallel Services only) | Y | ||||
CLUGROUPNAME
|
Cluster group name (Cluster Services only) | Y | ||||
CLULOCKNAME
|
Cluster lock structure name (Cluster Services only) | Y | ||||
CLULOCKSIZE
|
Cluster lock area size (Parallel Services only) | Y | ||||
CLUCACHETYPE |
Cluster cache data storage type (Parallel Services only) | Y | ||||
CLUSTER
|
Adabas cluster session control | Y | ||||
CT
|
Command time limit (seconds) | Y | ||||
DATACACHE
|
Controller caching control for the Data Storage component | Y | ||||
DIRRATIO/ELEMENTRATIO
|
Ratio of directory entries to data elements in a cluster cache structure/area | Y | ||||
DBID
|
Database ID (physical) | Y | ||||
DEVICE
|
Device type of the first ASSO extent | Y | ||||
DSF
|
Delta Save Facility control | Y | ||||
DSFEX1
|
Delta Save Facility user exit | Y | ||||
DTP
|
Distributed transaction processing control | Y | Y | |||
DUALCLD
|
Dual command log device | Y | ||||
DUALCLS
|
Dual command log size (blocks) | Y | ||||
DUALPLD
|
Dual protection log device | Y | ||||
DUALPLS
|
Dual protection log size (blocks) | Y | ||||
FASTPATH
|
Adabas Fastpath control | Y | ||||
FMXIO
|
Limit parallel I/O operations by LFIOP flush processing | Y | ||||
FORCE
|
Overwrite IDTE | Y | ||||
HEXnn
|
Hyperdescriptor exit(s) | Y | ||||
IDTNAME
|
BS2000 ID table name | Y | ||||
IGNDIB
|
Ignore DIB entry | Y | ||||
IGNDTP
|
Ignore distributed transaction processing area (Work part 4) | Y | ||||
INTNAS
|
Interval between nucleus statistic checkpoints (SYNS 60) | Y | ||||
LBP
|
Length of buffer pool | Y | ||||
LCP
|
Length of security pool | Y | ||||
LDEUQP
|
Length of unique (UQ) descriptor pool | Y | ||||
LDTP
|
Length of distributed transaction processing area (Work part 4) | Y | ||||
LFIOP
|
Length of asynchronous flush pool | Y | ||||
LFP
|
Length of internal format buffer pool | Y | ||||
LI
|
Length of ISN list table (TBI) | Y | ||||
LOCAL
|
Nucleus (cluster) unknown to the network | Y | ||||
LOGCB
|
Log control block | Y | ||||
LOGCLEX
|
Log command log extension (CLEX) | Y | ||||
LOGFB
|
Log format buffer | Y | ||||
LOGGING
|
Logging of Adabas commands | Y | ||||
LOGIB
|
Log ISN buffer | Y | ||||
LOGIO
|
Log I/O activity | Y | ||||
LOGRB
|
Log record buffer | Y | ||||
LOGSB
|
Log search buffer | Y | ||||
LOGSIZE
|
Maximum command log size | Y | ||||
LOGUX
|
Log user exit B data | Y | ||||
LOGVB
|
Log value buffer | Y | ||||
LP
|
Length of data protection area (Work part 1) |
Y | ||||
LQ
|
Length of sequential command table | Y | ||||
LRDP
|
Length of the deferred caching redo pool in cluster environments. | Y | ||||
LS
|
Length of sort area | Y | ||||
LU
|
Length of intermediate user buffer | Y | ||||
LWKP2
|
Length of ISN list processing area (Work part 2) | Y | ||||
LWP
|
Length of Adabas work pool | Y | ||||
MODE
|
Mode of operation | Y | ||||
MSGBUF
|
Size of the message buffer | Y | ||||
MXMSG
|
Maximum message reply time between cluster nuclei in Cluster Services or Parallel Services | Y | ||||
MXTNA
|
Maximum inactivity time limit override for a user | Y | ||||
MXTSX
|
Maximum Sx execution time limit override for a user | Y | ||||
MXTT
|
Maximum transaction time limit override for a user | Y | ||||
NAB
|
Number of attached buffers | Y | ||||
NC
|
Number of command queue elements | Y | ||||
NCLOG
|
Number of command logs | Y | ||||
NH
|
Number of hold queue elements | Y | ||||
NISNHQ
|
Number of ISNs in hold queue for user | Y | ||||
NONDES
|
Non-descriptor searches | Y | ||||
NPLOG
|
Number of protection logs | Y | ||||
NQCID
|
Number of active command IDs per user | Y | ||||
NSISN
|
Number of ISNs per ISN table element | Y | ||||
NT
|
Number of threads | Y | ||||
NU
|
Number of user queue elements | Y | ||||
NUCID
|
Custer nucleus ID | Y | ||||
OPENRQ
|
Open command required | Y | ||||
PLOGDEV
|
Multiple protection log device type | Y | ||||
PLOGRQ
|
Protection log required | Y | ||||
PLOGSIZE
|
Multiple protection log size (blocks) | Y | ||||
PREFETCH**
|
Prefetch/multifetch feature control (see note below) | Y | ||||
PREFICMD**
|
Include command from prefetch/multifetch (see note below) | Y | ||||
PREFIFIL**
|
Include file from prefetch/multifetch (see note below) | Y | ||||
PREFNREC**
|
Multifetch record count (see note below) | Y | ||||
PREFSBL**
|
Prefetch single buffer length (see note below) | Y | ||||
PREFTBL**
|
Prefetch total buffer length (see note below) | Y | ||||
PREFXCMD**
|
Exclude command from prefetch/multifetch (see note below) | Y | ||||
PREFXFIL**
|
Exclude file from prefetch/multifetch (see note below) | Y | ||||
QBLKSIZE
|
Sequential data set/file block size (optimized by ADAIOR) | Y | ||||
READONLY
|
Read-only session control | Y | ||||
REVIEW
|
Adabas Review control | Y | ||||
SMGT
|
Error handling (PIN) facility control | Y | ||||
SORTCACHE
|
Controller caching control for the Adabas sort area component | Y | ||||
SPT
|
Adabas triggers and stored procedures control | Y | ||||
SVC
|
SVC number | Y | ||||
TCPIP
|
TCP/IP access control | Y | ||||
TCPURL
|
TCP/IP universal resource locator (URL) | Y | ||||
TEMPCACHE
|
Controller caching for the Adabas temp area component | Y | ||||
TFLUSH
|
Synchronous buffer flush time | Y | ||||
TLSCMD
|
Time limit for S1, S2, and S4 complex searches (seconds) | Y | ||||
TNAA
|
Non-activity time limit (access-only users) | Y | ||||
TNAE
|
Non-activity time limit (ET logic users) | Y | ||||
TNAX
|
Non-activity time limit (exclusive update users) | Y | ||||
TT
|
Transaction time limit | Y | ||||
UEXn
|
User exits: 1, 3, 4, 5, 8 | Y | ||||
UEXn
|
User exits: 2, 12 | |||||
UEXn
|
User exits: 6, 9 (for utilities) | Y | ||||
UTIONLY
|
Utilities-only session | Y | ||||
VISTA
|
Adabas Vista control | Y | ||||
WORKCACHE
|
Controller caching for the Adabas work area component | Y |
Notes:
DTP=TM or DTP=RM. It does support
DTP=ET, which is a parameter setting that must be the
same for all cluster nuclei and cannot be changed (global, fixed).
PROGRAM=USER) making Adabas calls.
They have no effect when specified for an Adabas nucleus.
When specifying ADARUN session parameters for Adabas Parallel Services cluster nuclei:
ensure that the correct program to be executed is specified (PROG=ADANUC); and
determine which setting is applicable for the
SVC parameter for the session.
The CLOGMRG, CLUSTER,
CLUCACHESIZE, CLUCACHETYPE,
CLULOCKSIZE, DIRRATIO /
ELEMENTRATIO, LRDP, and
NUCID parameters are used by the Adabas Parallel
Services cluster nucleus and its environment.
If protection logs and/or command logs are used in a cluster environment, they must be dual or multiple logs and all nuclei must use them. All cluster nuclei must have the same PLOGRQ setting.
The remaining Adabas cluster nucleus parameters are the same as those of a standard Adabas nucleus. For more information, read Adabas operations documentation.