cDSCHEDULE Automated Job Submission System Guide and Reference �ManualD

SCHEDULE
Automated Job Submission System
Guide and Reference Manual

o T Y

2.3.2 Configuring a Workgroup

HThe Workgroup is configured using either the EnterpriseSCHEDULE Windows DClient or using commands. For a description of the commands used to Mconfigure the workgroup, see Chapter 12.5

2.3.2.1 Execution mode

EThe job's execution mode determines how jobs assigned to a workgroup #will be processed. The choices are:

Run on the least busy node in the workgroup - This selection will G cause the scheduler to choose the node that is currently running the F least number of jobs based on it's power rating, job limit and Vote count.J
Run on the first accessible node in the workgroup - This selection E will cause the scheduler to find the node that is first accessible * based on it's order in the members list.G
Run on all nodes in a round robin fashion - This selection will G run jobs throughout the Workgroup in a round robin fashion. In other H words, the first job will run on the first node in the Workgroup, the * second will run on the second node, etc.I
Run on ALL nodes in the workgroup - This selection will cause the H scheduler to run every job on every node. Keep in mind that this will G cause multiple runs of each job. Once a job runs with the run-on-all I workgroup type, all downstream jobs are required to be controlled by a I workgroup that is also of type run-on-all. In addition, the node where B each job instance is running must be a member of that workgroup.B
No jobs can run in this workgroup - No jobs will run in the Workgroup

2.3.2.2 Database mode

*The settings for the Database options are:

The database is shared using a fully clustered system - The I database is already shared using a fully clustered system and will not use any Workgroup replication.B
The database is shared using a limited cluster system - The G database is already shared using a limited clustered system and will $ not use any Workgroup replication.B
The database is distributed between nodes - The database is J distributed between nodes and will be heterogeneous. Each node may have different database entries.E
The database is distributed and replicated between nodes - The ? database is distributed across nodes and all entries will be E replicated. Every node in the databse will have identical folders, ! jobs, calendars, variables etc.H
Workgroup not used for database control - The data stays in it's 8 default location and is not distributed or replicated.

2.3.3 Directories

GThe SCHEDULE System supports two levels of directories, a top level and oneDsublevel. Any number of each variety can exist. The basic directory formats are: $ &
Directory Description
[000000] Root directory
[toplvl] Top level directory
[toplvl.sublvl] Sublevel directory
[DEFAULT] Last resort defaults

Directory	Description
[000000]	Root directory
[toplvl]	Top level directory
[toplvl.sublvl]	Sublevel directory
[DEFAULT]	Last resort defaults

-The default directory is the user's username.FEach directory can contain three types of entries. 1) Job entries. 2) GCalender entries. 3) Directory entries. All the entries are owned by a UIC code and have standardEVMS RWED protection settings possible. The job and directory entries Bcan also have an Access Control List (ACL) specified. ⁷

N Note

⁷ For a full explanation of Access J Control Lists please refer to the DEC VMS documentation set.

o V T Y

Contents

Index