c D SCHEDULE Automated Job Submission System Guide and Reference Manual D SCHEDULE
Automated Job Submission System
Guide and Reference Manual
9 Appendix E
Record layout
[ E.1 Control file fields
G The CONTROL file contains several record types. The primary one is the ? SDE record which is documented below. The others are directory C definitions and list specification, and these are not suitable for
external use.
B Whenever a new job is created a new CONTROL record is made. A new @ record is first loaded with a set of default values before user H specified values are loaded. If the job entry called "[DEFAULT]DEFAULT" E is found, that is used as the initial default. Otherwise a series of hard values are used.
E The general description of all fields in the CONTROL file are listed below:
J - SUBMIT OPTIONS: All the various submit parameters below only apply D to the batch job that is created when execution is requested. They A are not used to control any other facet of the SCHEDULE F system. For more information on what these do please refer to the VMS DCL SUBMIT command description.
I - USER: All automated submit operations are done on the behalf of this 4 user. Only a user with CMKRNL can change this field.H
- CLI: The CLI field specifies which command language interpreter A that will be enabled for the batch job. The default value is DCL. E
- NODE: The NODE field specifies on which node this job is to > execute. Inside a cluster any node that is a member of the ; same cluster is allowed to submit and execute the job. This H parameter is very useful in a "Satellite/Central database" 1 C Satellite/Central database. 2) Peer-to-Peer systems. )
network D configuration. In this network topology the job will execute on the F remote node. All the control data and DCL commands will come from the central database. D - CHARACTERISTIC: The CHARACTERISTIC field is a single known 9 characteristic name that is to be associated with the H job when it is submitted. Characteristics are arbitrary attributes that G can be assigned to job queues. A job will only execute in a queue with matching characteristics. L
- CPUTIME: The CPUTIME field specifies the maximum CPU time that the 3 job will execute. If the batch queue also has a D CPU time limit the smaller of the two is used. If none is specified D this implies NO LIMIT. The field contains a standard VMS delta date specification.H
- FLAGS: The submit flags are HOLD, KEEP, PRINT and DELETE. These C flags are the standard VMS batch flags and they have the following meanings.
J - HOLD - Place the job in the PENDING state on entering into the batch queue.C
- KEEP - Keep the log file on line after the job finishes. >
- PRINT - Print the log file after the job completes.D
- DELETE - Delete the job temporary COM file on completion.
B - LOG_FILE: The LOG FILE field specifies where to place the F job log file that is created during job execution. Any file name that F is provided is parsed and a default of "SYS$SYSLOGIN:job_name.LOG" is & used to complete any missing elements.C
- QUEUE: The QUEUE field defines what batch queue the job is D to be placed in for execution. This can be either a generic or node F specific execution queue. The default if none has been defined either & implicitly or explicitly is SYS$BATCH.E
- PRINTER: The PRINTER field defines what print queue that the F log file, if it is to be printed, is to be inserted into. The default F if none has been defined either implicitly or explicitly is SYS$PRINT.J
- NAME: The NAME field defines the job name that is to be assigned ? this job when it is in the batch execution queue. This name H will only appear when you use the VMS command SHOW ENTRY or SHOW QUEUE. 9 By default this name is the same as the job entry in the SCHEDULE database. L
- PARAMETER: The PARAMETER fields (P1 through P8) are just character I data that is passed into the batch jobs as DCL symbols P1, P2 ... P8. A They can be used in what ever fashion is needed by the developer. G
- PRIORITY: The PRIORITY field specifies what queue priority to L assign this job when it is inserted into the batch queue for execution. 1 A non-privileged user is limited to a maximum 8 value that is controlled by the AUTHORIZE program and/or B the SYSGEN parameter MAXQUEPRI. The priority is NOT the execution C priority but strictly the priority used to order jobs in the batch 0 execution queue while they are waiting to start.K
- ACCOUNT: The ACCOUNT field specifies what account value to assign G the job when it is submitted for execution. This parameter is only > useful if you have an application that can use this value.J
- WSDEFAULT: The WSDEFAULT field defines for a batch job a working L set default, the default number of physical pages that the job can use. If the queue on which the C job executes has a nonzero default working set, the smaller of the F specified job and queue values is used. If the queue on which the job F executes has a working set default of 0, the smaller of the specified F job and UAF values (value established in the user authorization file) H is used. If you specify the value 0 or NONE, the specified queue or UAF A value is used. Working set default values must range between the numbers specified by the SYSGEN $ parameters PQL_MWSDEFAULT and WSMAX.J
- WSEXTENT: The WSEXTENT field defines for the batch job a working L set extent, the maximum amount of physical memory that the job can use. The job uses F the maximum amount of physical memory only when the system has excess A free pages. If the queue on which the job executes has a nonzero F working set extent, the smaller of the specified job and queue values B is used. If the queue on which the job executes has a working set H extent of 0, the smaller of the specified job and the value established G in the user authorization file (UAF) is used. If you specify the value H 0 or NONE, the specified queue or UAF value is used. Working set extent % values must range between the numbers ; specified by the SYSGEN parameters PQL_MWSEXTENT and WSMAX. L
- WSQUOTA: The WSQUOTA field defines for the batch job a working set L quota, the amount of physical memory that the job is guaranteed. If the queue on which E the job executes has a nonzero working set quota, the smaller of the F specified job and queue values is used. If the queue on which the job H executes has a working set quota of 0, the smaller of the specified job G or the value established in the user authorization file (UAF) is used. H If you specify the value 0 or NONE, the specified queue or UAF value is H used. Working set quota values must range between the numbers specified by the SYSGEN parameters PQL_MWSQUOTA and WSMAX.
F - NOTIFICATION OPTIONS: Whenever a notification event occurs the H information regarding the event is collected and place in a temporary . COM file. This file is then submitted to the F notification batch queue indicated in the system parameter file. This F notification procedure 2 parameter in the parameter file.) C is capable of delivering two types of messages. One is a broadcast G message to the user (if still logged in). The other is a mail message. G The user name or names that are specified can be logical names that in G turn provide a distribution list for all the various messages.
The 3 various notification events are detailed below: "." ( - SUBMITTED: This field indicates H that the user(s) is to receive a message when the job has been submited for execution.&
- STARTED: This field indicates F that the user(s) is to receive a message when the job actually starts
execution.&
- DELAYED: This field indicates H that the user(s) is to receive a message when the job has been delayed. C A job is considered delayed if it stays in a "wait for" scheduling F queue longer then the AFTER_DELAYED time. In all the other scheduling E queues a job is considered delayed if it stays there longer then the - value indicated by the STUCK_AFTER parameter. (
- COMPLETED: This field indicates D that the user(s) is to receive a message when the job has completed
successfully. %
- FAILED: This field indicates H that the user(s) is to receive a message when the job fails. During the B processing of scheduling requests various problems can arise (for G example invalid calendar name, etc.). All such events are also trigger failure notification events.,
- AFTER_DELAYED: This field indicates 5 the interval after which a job is considered delayed.
H - RESOURCE REQUIREMENTS: Two types of resource requirements can be 8 imposed on a job. Both requirements must be met before H a job is submitted for execution. Two sets of resource used information 7 are collected. These parameters are detailed below: "."
H - DISK_SPACE: The DISK SPACE field indicates how much disk space 7 must be available on a certain disk before this job F is to start. The job will wait in the scheduling queue "wait for disk C space" until such time that this amount of space becomes available. L
- FILE: The FILE field indicates that a certain file must be present 4 for the job to proceed. The job will wait in the H scheduling queue "wait for file" until such time that this file becomes
available.L
- MEASURED: As each job completes the usage information is collected , and add in to the average usage statics. . The following four numbers are maintained. "-"
L - CPU TIME: The average number of seconds for CPU time used by the
job.N
- ELAPSE TIME: The average number of seconds of wall clock time used by the job. N
- IO COUNT: The average number of I/O's (buffer I/O plus direct I/O) that the job uses.H
- COUNT: The number of runs that this average is computed over.
G - ESTIMATED: Each job can be provided an estimated usage value. I These numbers are not changed by the system and are there to provided G an additional slot to store planned or estimated values. The following fields are provided. "-"
6 - CPU TIME: An estimated CPU time in seconds.<
- ELAPSE TIME: An estimated elapse time in seconds.8
- IO COUNT: An estimated I/O count for the job.
A - GENERAL INFORMATION: Each job contains a group of general
information ? describing the job, access to job and when the job is to start.
H - COMMENT: The COMMENT field is just a text field into which any data can be inserted. I
- NEXT_SUBMIT: The NEXT_SUBMIT data is the next time the job will ? automatically start. Any that job that is to start (without H being manually started) must have a start time set. Once the start time E has occurred and the job submitted for execution then the reschedule D parameters are used to determine how to reset this date to the next start time. C
- LAST_START: The LAST_START time is an output field that is G updated whenever the job has been submitted and is starting to execute. E
- LAST_SUBMIT: The LAST_SUBMIT time is an output field that is G updated whenever the job is submitted into a batch queue for execution. H
- LAST_COMPLETED: The LAST_COMPLETED time is an output field that ? is updated whenever a job completes, either success or failure. K
- RESTART_COUNT: The RESTART_COUNT field indicates how many times a C job is allowed to be automatically restarted. A non-zero value A indicates that the job can be restarted on failure (error and H severe error conditions) and how many times it is allowed. Any initiate F jobs that are specified are not activated until a job has failed this F number of times. Detected job failure events include such things as a H node going down, a job explicitly deleted out of the queue or a program failure.G
- UP_COUNT: The UP_COUNT field is incremented each time a job is submitted for execution.H
- DOWN_COUNT: The DOWN_COUNT field is decremented each time a job G submitted for execution. When a value of zero is reached all automatic B job submissions terminate. A value of INFINITE or -1 indicates an % unlimited number of runs are allowed. H
- OWNER: The owner field indicates the UIC that controls the job. / This can be either an IDENTIFIER or a UIC code. H
- PROTECTION: The protection field is a bitmap that grants access G to the system, owner, group and world classes. The protection bits are used as follows:
/ - D - Delete access allowed to the job .
- W - Write access allowed to the job-
- R - Read access allowed to the job >
- E - submit, initiate or prerequisite access allowed
1
The access modes are nested as follows:
/ N Note
œ 1 Please refer to an earlier section describing network topologies. Two distinct topologies are supported. 1 2 The procedure SCHEDULE_LIBRARY:SCHEDULE_NOTIFY.COM formats the event information and actually transmits the event notification message. If you need to modify this procedure make a copy of it and change the SCHEDULE_NOTIFY |