AU2004222721A1 - System for Scheduling and Monitoring Computer Processes - Google Patents

Description

0 0 c( o SYSTEM FOR SCHEDULING AND MONITORING COMPUTER

SPROCESSES

ci Cl Background of the Invention o Field of the Invention This invention relates to the scheduling and monitoring of computer processes. The invention is more particularly related to the submission of jobs for execution. The invention is still further related to the submission, scheduling and monitoring of jobs across multiple networked computer platforms (nodes) and the provision of a common interface for programs submitting to the jobs.

The invention is still further related to a job scheduler that maintains local job repositories having detailed job histories for each node and a central job repository maintaining detailed job history across an enterprise. The invention is yet further related to the provision of a scheduling agent on each computer platform to start execution of each job submitted.

Discussion of the Background Modern computer systems are utilized for a wide range of tasks. Many tasks are simple and are executed in real time. However, some tasks require long execution times, or must be performed at various intervals or at inconvenient times (when a system where a task is running has a light tasking load, early morning or weekend hours, for example).

Basic scheduling devices have been utilized to run PCT/US99/31024 -2

O

o certain programs or jobs at various intervals or at o specified run times. However these systems do not provide adequate service or integrate seamlessly into any specific product line, nor provide appropriate Ci 5 service between multiple computing platforms in a networked environment.

ci SUMMARY OF' THE INVENTION Accordingly, it is an object of this invention to provide a job scheduling apparatus for scheduling jobs.

It is another object of the present invention to provide a job scheduling apparatus that provides job scheduling services across multiple computing platforms, and control over the execution of a job submitted; It is yet another object of the present invention to provide a scheduling agent on each respective node of a computer network for accepting and managing jobs submitted to the respective node; It is another object of this invention to provide a command line that may be utilized to determine job status and issue job control commands to jobs executing on a node in an enterprise.

It is another object of this invention to provide a seamless job scheduling device for plural software products having a common format for submission and scheduling of jobs, and to provide consistent application programming interfaces to the software products utilizing the job scheduling device.

It is yet another object of this invention to provide a single job scheduling and administrative tool for all POEMS enabled products (Platinum point products, for example) under a common application programming interface that specifically and efficiently WO 00/38033 PCT7US99/31024 0- 3 targets job scheduling requirements of the products.

0- It is yet another object of the present invention to provide a common graphical user interface (Microsoft Foundation Class (MFC), for example) to schedule and list all jobs and common APIs used by a GUI component C and the agent.

o BRIEF DESCRIPTION OF THE DRAWINGS 0 A more complete appreciation of the invention and many of the attendant advantages thereof will be readily obtained as the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawings, wherein: Figure 1 is a block diagram illustrating an implementation of the present invention providing a job scheduling and administrative tool for POEMS enabled products under a common application programming interface.

Figure 2 is a block diagram of a Lightweight Enterprise Scheduler (LES) Agent that coordinates execution and job history submissions for a node; Figure 3 is a flow diagram illustrating an RTserver utilized to communicate messages to Receive modules; Figure 4 is a flow diagram illustrating an RTserver utilized to communicate messages to Receive modules subscribed to a specific node; Figure 5 is a flow diagram illustrating LES agent receipt of Platinum Enterprise Communicator

(PEC)

messages; Figure 6 is a flow diagram illustrating an LES agent sending point product parameters to a point product process via PEC; WO 00/38033 PCT/US99/31024 0- 4 0 Figure 7A is a flow chart illustrating startup of C-i an LES agent; Figure 7B is a flow chart illustrating the processes of an LES Agent; C' 5 Figure 8 is a block diagram illustrating an LES CA directory structure; C< Figure 9 is a block diagram of a Platinum Security o Model illustrating user identification requirements of 0 applications and point products; Figure 10 is an illustration of a POEMS Scheduling Service Job Scheduling window; Figure 11 is an illustration of a Point Product Property sheet; Figure 12 is an illustration of a pull-down calendar on a Point Product Property sheet; Figure 13 is an illustration of an example job scheduled to run via the Point Product Property sheet; Figure 14 is an illustration of a second example of a job scheduled to run via the Point Product Property sheet; Figure 15 is an illustration providing an example of a job scheduled to run every day; Figure 16 is an illustration that provides an example of- a job scheduled to run every week on a specific day; Figure 17 is an illustration providing an example of a job scheduled to run on a selected specific date of each month; Figure 18 is an illustration that provides an example of a job scheduled to run once a year on a specific date; Figure 19 is an illustration of a Notification Script window; WO 00/38033 PCTIUS99!3] 024 Figure 20 is an illustration of a Calendar C] Selection window; Figure 21 is an illustration of a Create Calendar window utilized to create a calendar having selected

C]

5 dates for running jobs associated with the calendar; C' Figure 22 is an illustration of an existing C calendar brought up for editing; o Figure 23 is an illustration of a Strategy C\ Scheduling window utilized for creating strategy windows and selecting strategy windows for editing and deleting; Figure 24 is an illustration of a Point Product Property sheet invoked by clicking Create from the Strategy Scheduling window; Figure 25 is an illustration of a Delete Schedule window having a prompt for identifying a clean-up script upon deletion of a schedule; Figure 26 is an illustration of an Insert Object window; Figure 27 is an illustration of a General tab of a properties window; Figure 28 is an illustration of a Representation tab of a properties window; Figure 29 is an illustration of a Select Intellicon bitmap utilized for selecting icon graphics for representing an object; Figure 30 is an illustration of an example in an Explorer view for the display of labels for a jobs resource object; Figure 31 is an illustration of a Job Repository tab of a properties window; Figure 32 is an illustration of a hierarchy of folders of a jobs resource object; WO 00/38033 PCT,/JS99!31024 0

O

O Figure 33 is an illustration of a listing ot ]obs 0 in All Jobs folder; Figure 34 is an illustration of a job in an All Jobs Any Status folder; c 5 Figure 35 is. an illustration of an example of a Ci Job's Run history;

CA

Figure 36 is an illustration of a listing of jobs Sin an All Jobs Any Status folder under a specific node; c o Figure 37 is an illustration of Property Page tabs that are available for jobs; Figure 38 is an illustration of a Command Tab property page; Figure 39 is an illustration of a Databases Tab property page; Figure 40 is an illustration of a Job Scheduling Tab property page; Figure 41 is an illustration of a Parameters Tab property page; Figure 42 is an illustration of a General Tab property page containing information about a job run; Figure 43 is an illustration of a Run Stats Tab property page that displays product specific information about a job run; Figure- 44 is an illustration of a General Tab property page on the text of a group in a Jobs resources object; Figure 45 is an illustration of a view of a Log File column in a Director Explorer view; Figure 46 is an illustration of utilization of a popup menu for viewing a Log File; Figure 47 is an illustration of a Log File viewer displayed for a specific job run; Figure 48 is an illustration of a popup menu utilized to delete a folder of jobs; WO 00/38033 PCT/US99/31024 7 Figure 49 is an illustration of a Delete Jobs Cs] window; Figure 50 is an illustration of a Status and Results window;

CA

Figure 51 is an illustration of utilization of a popup menu to delete a single job; Figure 52 is an illustration of a Delete Jobs o window; Figure 53 is an illustration of a Status and Results window for the deletion of a single job; Figure 54 is an illustration of utilization of a popup window to evoke a rerun of multiple jobs at a folder level; Figure 55 is an illustration of a Rerun Jobs popup window; Figure 56 is an illustration of a Status and Results window for the rerun of jobs in a Jobs folder; Figure 57 is an illustration of a utilization of a popup window to rerun a single job; Figure 58 is an illustration of a Rerun Jobs window for an individual job; Figure 59 is an illustration of the utilization of a popup window to cancel running jobs; Figure 60 is an illustration of a Cancel Runs popup window; Figure 61 is an illustration of the utilization of a popup window to invoke a Progress Monitor on a running job; Figure 62 is an illustration of a Progress Monitor utilized by the present invention to show phase, overall prog and other specific information about a job; Figure 63 is an illustration of a Director Service Manager Monitor; and WO 00/38033 PCT/US99/31024 8 0 Figure 64 is an illustration of a right side of a 0 Service Manager Monitor displaying information for programs residing on a node.

DESCRIPTION OF THE PREFERRED EMBODIMENTS (C Referring now to the drawings, wherein like Ci reference numerals designate identical or corresponding o parts throughout the several views, and more 0 particularly to Figure 1 thereof, there is illustrated an implementation of the present invention providing a job scheduling and administrative tool for POEMS enabled products (Platinum point products, for example) under a common application programming interface.

In Fig. 1, a suite of point products 100 (point products each including a point product GUI 110, are linked together with various POEMS enabling application programming interfaces (API's), and a Platinum Enterprise Communicator (PEC) 160. At least one Lightweight Enterprise Scheduler (LES) agent 170 is installed on a node. The point product GUI 110 provides a user interface to the point product 100 and communicates user selections by sending job data to a GUI API 115.

The GUI API 115 incorporates LES provided property pages, performs data checks and job allocation, and calls relevant scheduling functions in the job scheduling API 130 according to the job parameters.

The GUI API 115 is used by the point product GUI 110.

The job scheduling API 130 performs scheduling services, allocates and defines jobs submitted and distributes jobs by initiating a PEC message to an LES agent on a selected node to run the job (LES node 1, 170, for example). Inter-process communications are performed via the PEC 160. The Job Scheduling API is WO 00138033 PCT/US99/31024 -9 o preferably used by the point product GUI 110.

o The LES agent 170 initiates processing of the job as requested by the Job Scheduling API 130. The present invention may be implemented across multiple C 5 computing platforms and/or nodes, and an LES agent is CA installed on each node on which jobs run (LES agents C SEach LES Agent updates a local job repository Is0 Sthat maintains job information on each job submitted to it's respective node. The local job repositories are installed on each node where jobs run and may be implemented using SQL*Anywhere (a database product having a small footprint, for efficiency of resources).

Other SQL products and various database/repository configurations may be substituted.

A central job repository 190 maintains a super set of job data on all jobs submitted to each node in the system. The central job repository is maintained by a job data management API 140, and is implemented in a Relational Database Management System (RDBMS), Dex, or other data repository or management system.

The job data management API 140 selects, inserts, updates, or deletes jobs and job histories. Both local and central job repositories are updated by job data management API. Also, the job data management API 140 is used internally by the job scheduler and the LES agent. The job scheduling API reformats data into PEC message format.

A location of the central job repository 190 is determined by RDEX (Relational Data Exchange). If no DEX (Data Exchange) information is available, the SDM (Software Delivery Manager) install prompts are utilized for table location information.

Configuration files for set up of the POEMS W-O 00/38033 PCT/US99/3 1024 0 environment, PEC 160, LES agents, etc. are maintained o on a same node as the central job repository, ea--.

client machine, and on each node where jobs run.

Each of the Central and local job reposJ tories are automatically updated by LBS. The point- products themselves only interact with the API 's and are not concerned with the details of either type cf o repository.

o A point product API 120 is provided to allow access by each point product to determine job status, job logfile, perform setup and cancel functions (jobs), update the job logfile, and retrieve job parameters as the j ob runs. In the POEMS configuration, this APT is utilized by the point product executable (point product 100, for example).

An LBS command line 155 provides command line access to job administration functions (by communicating with a job administration API 150), including determine job status, and setup, cancel, or update a job logfile and retrieve job parameters.

The job administration API 150 deletes, cancels, reruns, or copies existing jobs. This API is used internally by the LES command line 155 and a job folder 185 (containing a list of jobs, that may be viewed in various formats).

Each of the point product API 120, Job Scheduling API 130, Job Data Management API 140, and Job Administration API 150, communicate as needed with any of LBS agents installed on nodes 1. .n -via PEC messages.

The PEC messages are provided in a f ormat to communicate 'between LBS enabled workstations and the various APIs.

WO 00/38033 PCT/US99/31 024 0 0 0 0 A calendaring API 165 is provided to manage O calendar functions and is used internally by other API's and the point product GUI.

The POEMS enabled point products and LES agents C 5 installed on nodes of a computing network or system Cl provide common job scheduling services, hereinafter Cl referred to as the Lightweight Enterprise Scheduler S(LES). The LES allows a user to schedule a new job, Sdelete a job, cancel a running job, and re-run a job.

A sample LES job flow is provided in the following steps: 1. Enter job parameters in the point product

GUI.

2. Press the "Finish/OK" button.

3. Job parameters are passed to the LES Job Distributor.

4. Parent job entry is placed in the central repository.

Job parameters are passed to the LES agent using PEC.

6. Job parameters are stored in the local Job Table. (If this step fails, an event/alarm is sent indicating that the job is not scheduled on the node.) 7. Child job is created and stored in the central repository.

8. Repeat steps 5-7 for each of selected nodes.

9. At the appropriate time, the job is started (which launches the point product executable).

Job parameters are obtained from the job table using a LES API.

11. A job progress message is sent out.

WO 00/38033 PCT/US99/31024 c,12 o 12. The job's progress displays on the console 0 using a progress monitor.

13. The job completes, and the LES agent places a record in the local Job History Table.

Cl 5 14. The LES agent places a record in the central C- Job History Table.

C-i 15. The LES Agent sends a "Job Complete" event.

SLES Agent Figure 2 is a block diagram illustrating the main components and communications of an LES agent (LES agent 170, for example). The LES agent consists of three main parts: 1. Communication Module 210 Sends and receives PEC messages from the Job Distribution API to the point product and from the point product to the Job Distribution

API.

2. Job Management Module 220 The LES kernel.

This part of the agent sets up the internal environment and memory, launches the job, runs the job, and manages job processes.

3. Data Management Module 230 Updates and deletes data from the Local and Central Repositories (180 and 190, respectively).

A low-level function API 240 is available to all agent modules. This low-level API handles all internal functions, file management, and messages.

Figure 3 is a flow diagram illustrating how PEC messages are normally routed. A send module 300 sends a message 320 having a destination encoded therein WO 00/38033 PCT/US99/31024 0

O

13 0 which is received by an RT server which sends copies of o the message 320a and 320b to each of Receive modules Cl 340a and 340b, respectively (LES agents, for example) To prevent the RT server from delivering the same Cl 5 job message to more than one agent, the LES agen C( registers as PTLES_<nodeaddress> datagroup. This (C registration identifies each LES agent with a unique datagroup so that messages may be routed to nodes of a o datagroup corresponding to the message.

Each LES Agent subscribes to the current node address (hostname) messages, so each sender should also specify which node will receive the message. Proper message routing is illustrated in Figure 4, which illustrates the RT server 310 passing the message 320 to Receivel module 345a, and Receive module 345b not receiving a copy. An RT Server runs on each machine where processes are managed, provides PEC routing to a correct destination.

Figure 5 is a flow diagram illustrating the flow of messages incoming to an LES agent 500. All incoming messages are received via a PEC communication API (PEC 160, for example, see Figure from any of a client process 520, point product agent 530, or other module communicating with an LES agent. Each message is routed via an RT server (510a, 510b, for example) to the LES agent 500.

Figure 6 illustrates reverse message traffic, from the LES agent 500 to the point product agent 530. The LES agent 500 sends point product parameters to the point product process (agent) 530 using PEC messages via the RT server 510.

Figure 7A is a flow chart illustrating the steps for LES agent startup. At step 705, LES configuration files (which maintain startup information on location WO 00/38033 PCT/US99/31024 0

O

14 O of repositories, information for LES to find out how to 0 set-up and operate) are loaded. At step 710, LES local job repository tables are created (local job repository 1 180, see Fig. 1, for example). Step 710 is bypassed C 5 if local job repositories are already present on the Cl node which the LES agent is being started.

C-i At step 715, the local job repository is Ssynchronized with the central job repository 190. For example, The synchronization process updates the central job repository to reflect each entry in the local job repository, this process assures that the central job repository maintains records for each job across the enterprise.

At step 720, a history row is added for all expired jobs (both central and local). Information on expired jobs is maintained for historical purposes.

At step 725. PEC call back functions are initialized. The PEC callback functions provide the appropriate API interface allowing communication with the PEC 160.

At step 730, a synchronization timeout is computed and setup. The synchronization timeout is utilized to control how long to wait to connect to the other repository.

As illustrated in Figure 7B, once the LES agent is started, it begins performing receipt and startup of job processes selected to be run on the node on which the LES is installed. At step 750, the LES agent computes and performs setup of a next job to run, and then enters a wait loop 760, where the LES agent waits for one of a PEC message, Synchronization timeout, and a timeout on a next job to run.

When the wait loop times out on a next job to run, step 770 is performed, which includes running the WO 00/38033 PCT/US99/31024 0 0 15 0 current job (next job that timedout) updating a start 0 status of the job, and creating a run history row (updating and creating actions are performed in both local and central repositories).

C 5 When the loop times out on a synchronization, a Central/Local synchronization program is executed Cl (synchronizes Local with central, and central with local).

0 When the loop receives a PEC message, the LES agent performs the action contained in the PEC message, step 780. PEC messages received by the LES agent include messages to run a job, cancel a job, update job status, request job parameters (Job parameters are information that a product would need, user name, db name, set by point product and stored in the LES API in an area of the LES db where point products can use for whatever they want).

Upon completion of either the job timeout, synchronization timeout, or PEC message action (steps 770, 790, and 780, respectively), the LES agent computes and sets up a next job to run (repeating step 750), and re-enters the wait loop 760.

When the POEMS Scheduling Service is utilized to add a job,. the point product GUI 110 calls the Job Scheduling API 130 to submit the job.

The API performs the following steps: 1. Saves the job in the central repository.

2. Checks to see if the agent is present.

3. If the agent is running, sends a PEC message (including the job) to the agent.

The Agent performs the following steps: 1. Receives the PEC message.

WO 00/38033 PCT/US99/31024 1-1

O

o 2. Saves the job in the local repository.

O 3. Checks when to run the job.

Cl 4. Launches the command line.

c- 5 The LES agent is maintained in an hierarchically organized directory structure, one example being Cl illustrated in Figure 8. A top level directory SPLATHOME 800 contains all relevant directories 0 containing files related to Platinum products (other products could alternatively be used). A POEMS root directory, $ptroot 810 maintains all directories and files related to enabling

POEMS.

A cfgref 830 directory maintains a configuration file configuration files 195, for example.

A shlib* 840 directory maintains all *.lib, and *.so files, which are shared libraries.

A bin 850 directory separately maintains each of a ptles* files 852, *.dll 854, and *sql.dll 856 files.

The ptles* files 852 include ptlesag, ptlestp, ptlestab, and other related files. The *.dll 854 maintains each of dynamic link libraries, and *sql.* 856 maintains LES queries for LES functions.

A log 860 directory maintains a ptlesag.log 865 logfile (that includes a diagnostic log, info about runs, and errors).

In one embodiment, as illustrated in Figure 9, the Platinum Security Model requires that many applications/point products be run as a specific user in order to succeed.

An AutoLogin API provides a way to run jobs as al6 specific operating system user without querying the user at run-time or hard-coding the UNIX user as part of the job.

WO 00/38033 PCT/US99/31024 0 17 0 0 To take advantage of this feature, an application 0 may utilize an PtLESSetJobCommand() function and pass a username and role to LES. If the username and role are set to NULL, LES runs the job as the default 5 platinum user (or other default user). Otherwise LES C' tries to retrieve the userid and password by calling CA' the AutoLogin API and passing the username and role as o Requested User and Requested Role, the operating system 0 as Resource Type, and the job instance name as Resource Instance.

Entering Login Information in AutoLogin In the AutoLogin embodiment, an administrator signs in as "root" to set up the LES/OS resources.

From the POEMS Director: 1. Select Tools*Security Admin- AutoLogin.

2. Log in as "root". If already logged in, "rtlogout.exell may be utilized before running Administrator AutoLogin.

3. The AutoLogin Administration window displays.

4. Right-click on the Agent's Requesting User folder to invoke a pop-up menu display.

Select Add Entry.

6. The Add Entry window displays.

7. Enter login information in the fields.

h) The Resource Type is set to OS.

i) The Resource Instance is the same instance name that the point product passes to the job.

WO 00/38033 PCT/US99/31024 8 o If the Point Product Clq *Uses AutoLogin and a valid user ID and a password are recorded in the Add Entry window, the LES C, aaent passes the ID and password strings and access is granted. Cl Cl Cl QUses AutoLogin but no user information is 0recorded in the Add Entry window, the LES agent 1l automatically checks for the UNIX user login. If the UNIX user is found, access is granted.

*,Uses AutoLogin and an invalid user ID and password are recorded in the Add Entry window, the LES agent defaults to PLATINUM (or other default) user.

fDoes not use AutoLogin, the LES agents defaults to the PLATINUM (or other default) user.

Command Line Interface The command line interface included in LES performs administrative functions, including: Delete a job with all its runs.

Cancel a job's run.

List all jobs.

List all jobs by product code, status, and/or node.

Rerun a job immediately.

The command line interface utilizes the following format: WO 00/38033 PCT/US99/31024 19 ptlescmd action] [-jobid] 1-c cleanup_process] run_number] product_code] status] [-b buffersize] node] Note: The -a option is for all actions; the -j option is for all actions except List; the others are optional.

Ptlescmd is the LES command line interface used to delete a job, cancel a job run, list jobs (all jobs, by product code, by status, and/or node), and rerun a job immediately.

Table 1 provides a listing of the ptlescmd command line supported options.

TABLE 1 F Command Line Options and Arguments Description -a [action] Error Executes a given action.

Action can have any of the following values: delete,cancel,list, rerun -j [jobid] Error (except The job ID of the job to be with "list" acted on. In one action embodiment, this is a required option (except with List action), -c (cleanupprocess] The name of the process to be executed after deleting a job. If a delete action is specified and no cleanup process name is given, then no cleanup will take place on UNIX after deleting the job. On NT this option will check the registry for an existing cleanup process and execute it after deleting the job.

i WO 00/38033 PCT/US99/31024 20 Command Line Options and Arguments i- If Omitted Description -r [run-number] Error In one embodiment, this number is required with Cancel action to define which run of the specified Sjob will be generated.

-p [productcode] Used only when product code is specified with List action, a list of all the jobs with this specific product code will be generated.

Used only when status is specified with List action, a list of all the jobs with this specific status will be generated.

-s [status] ITi -n [node] -b [buffersize] Default value is 1024 Used only when node is specified with List action, a list of all the jobs on this node will be generated.

This option is mainly needed when the user knows that the number of elements to be retrieved is large.

(>100,000). The default value is 1024(1K).

is; 1024 To use the LES command line, a user types the following: ptlescmd -a[action] -j[jobid] [-letter optionname] where: action is one of the following: delete, cancel, rerun, and list.

jobid is the identifier used for the job.

letter is the letter for one of the options listed in the Command Line Supported Options table.

WO 00/38033 PCT/US99/310: 21 optionname is the name of one of the options listed in the Command Line Supported Options table.

Table 2 provides a set of example command lines and a corresponding description.

TABLE 2 Command Line Description ptlescmd -a delete 123 Deletes job 123 and all its runs -c cleanup.exe then runs the process cleanup.exe ptlescmd -a delete -j 134 Deletes job 134 and all its runs ptlescmd -a list Lists all jobs ptlescmd -a list -n Lists all jobs on the node dimultra 'dimultra' ptlescmd -a list -p TSR Lists all jobs with the product code 'TSR' ptlescmd -a list -s running Lists all jobs and their running runs ptlescmd -a list -p TSR Lists all jobs of product code -s completed 'TSR' and their completed runs ptlescmd -a cancel -j 234 Cancels run 3 of job 234 -r 3 ptiescmd -a rerun -j 345 Reruns job 345 immediately Note: Valid status values include: Completed, Failed, Notstarted, Preempted, Running and Stopped.

LES jobs stay in the central and local repositories until the user deletes them. The user can delete, rerun, or cancel a running job at any time by rightclicking on the Poems director/Job folder 185 and selecting the appropriate option. A delete option removes the job and all the runs and history of the job.

24 WO 00/38033 PCTUS99/31024 0 -22 o The point product may provide a cleanup process o executable that removes all the point product files related to the job. This executable, if available, is run by the agent before removing the job and job 1 5 history.

LES provides jcb modification options including C' Rerun Job and Update Job. The Rerun Job option allows a user to rerun the job with a different start time.

o A PtLESUpdateJob() function allows updates for all GUI C 10 fields.

Context variables, such as %C(PLATHOME) or the environment variable %E(FUS_HOME) etc., which are resolved by the agent before running the job command.

The UNIX username is set from the GUI point product as part of the job. The agent does not need the password because the agent is running as root (the agent has the set_rootID bit) Context Variables Context variables are keyword identifiers that may be used on the command line. The LES agent converts these variables according to the context of the current job and job parameters passed. The LES agent inserts these context variables, forms the command line, and then executes the command line.

*Keyword Context Variable Substitution takes place from the current job.

*Environment Variable Substitution takes place from the environment where the agent is running.

*Filename Variable The agent uses this variable as the filename and replaces it with the contents of the file pointed to by the variable.

W1O 00/38033 PCTfUS99/31024 0 0 23 How Context Variables Are Specified: Keyword Context Variable: Specified as %C(context identifier is replaced by value)-e.g., my prog %-C(JOB_ID)

%C(USERNAME)

replaced with my prog 1234 manager where the current Job jobid 1234 and username manager.

Environment Variable: Specified as %E(variable name) my_prog %E(HOME)%E(LANoG), myprog/home/platinum en-us.

Filename Variable: Specified as %F(filename) my prog %F(/etc/platinum)%C(JOBID) %E(HOME)my_prog /home/dmc 1234 /home/platinum where the filename /etc/platinum contains the line /home/dmc.

Acceptable %C(conext variable name) syntax keywords: The syntax for the variables is %-C(KEYWORD). A user may insert a space on either side of the KEYWORD inside the parentheses. Table 3 provides a listing of sample keyword identifiers.

TABLE 3 Identifier Identiier Value Description JOB_ID-Env _JobRecord.joid Job ID GROUP_ID-Env _JobRecord.groupid Group HOSTNAME _JobRecord.node Hostname INSTANCE-EnV JobRecord.instance Instance name USERNAME _JobRecord.username User name WO 00/38033 PCT/US99/31024

O

c) 0 ci 24

OWNER

PASSWORD-Env CONNECT STR

JOBOWNER

OBJECT

PRODUCT CODE /*PRODUCT VERSION

/*DESCRIPTION

EMAIL ADDR

TYPE

C SCRIPT F SCRIPT

PLATHOME

USERROLE-Env JobRecord.owner I Owner I Owne JobRecord.password Password Pas wor _JobRecord.connectstring ConnectStrin _JobRecord.jobowne r _JobRecord.obj ect JobOwner Object name _JobRecord.productcode _JobRecord.productversion _JobRecord.description _JobRecord.notification email _JobRecord.type _JobRecord.completion script _JobRecord.failedscript Installation directory JobRecord.userrole Product Code Product Version Description E mail Address Type Completion script Failed script Install dir User role GUI API The GUI API is utilized by the point product 100 and the point product GUI 110 for a number of functions, including: *Allocate a job with the Job Scheduling API and override any default values as desired.

WO 00/38033 PCT/US99/31024

O

O

(C 25 O 'Base a GUI on class CptDynPropertySheet (which is derived from CpropertySheet) instead of using cN CpropertySheet directly. CptDynPropertySheet is exported by DnyPS.DLL In addition, the GUI API 5 may be used to submit a job without ever displaying the GUI window.

j *Use the GUI API to construct a SCptLESJobSchedulingPage (see Fig. 25, for Cl 10 example) object based on the job, and then add it (and any other pages) to the CptDynPropertySheet or CpropertySheet-derived property sheet.

*Use the GUI API to customize any job parameters that are displayed on the scheduling property page.

*Use the Job Scheduling API to set any required job parameters that cannot be set by controls on the scheduling property page. Examples of these include point product-specific job parameters set by calling PtLESSetJobParm.

*When a final OK indicates that the job should be submitted, use the GUI API to validate the data and submit the job.

Products that use the LES Job Scheduling page as a tab call its ValidateData() method before submitting the job. ValidateData() can be safely called multiple times and can be called in such a way that no message box appears to the user.

Table 4 provides sample code for invoking the GUI API to implement to the above discussed processes.

WO 00/38033 PCT/US99/31024 26 TABLE 4 Step Sample Code Allocate a job and override defaults Setup property sheet and pages Set product job info Customize the page's data When the sheet exits, validate the daza, and submit the job 1 PtLESJobHandle job PtLESJobAlloc() PtLESSetJobTZNode (job, PTLES_TZLOCAL); i CpropertySheet sheet; CptLESJobSchedulingPage LESpage (CptLESJobSchedulingPage::AsPage, job); Sheet.AddPage (LESpage); PtLESSetJobParam (job, key, value); LESpage.SetRunWhen(CPtLESJobSchedulingPa ge::RunImmediately); LESpage.SetRunWhenState(CPtLESJobSchedul ingPage::Disable) In the sheet's OK handler: if (LESpage.ValidateData(}) sheet.m_LESPage- >SubmitJob("nodename", groupID, TRUE).

Job Distribution

API

The job distribution appropriate machines called by the job data API dispatches information to (LES workstations), typically management API 140.

Job Scheduling API The job scheduling API 130 works with the point product GUI to define submitted jobs. Some of the functions available via the job scheduling API are required, some are optional, and some are internal functions that can be set by the user through the LES GUI. A value is set for baseline functions and any desired optional functions or functions that are defined by the GUI (defined as appropriate for a corresponding point WO 00/38033 PCT/US99/31024 -27 product) The following tables list functions by type and provide a brief description of each function. The baseline functions are utilized to provide basic information for the job scheduling processes, and are listed with a description in Table TABLE 5 Baseline Functions Function Description I User Name Node User name under which to run the job. In one embodiment, the scheduler must be run as root in order to use this function. If it is run as root and the User Name field is null, the scheduler will automatically mark the job as failed.

Indexed on central table. Specifies the node used to run this job.

i Command Product Code The command line to execute. Because of database VARCHAR size limitations, this may be split into 2 pieces by the API and reassembled to invoke.

Code identifier of the product used to create the job. This should be the code assigned and used by the PLATINUM PAR tracking system.

Version number of the product used to create the job.

Product Version The optional functions, listed and described in Table 6, are not required and are utilized (defined) as appropriate for a specific point product.

WO 00.38033 PCT/US99/31024 28 TABLE 6 Optional Functions Function GroupID ReTrwInterval Re try .Count N r e t rys (RecurFailThreshol d) IDescription Description Strategy group ID Use the function PtLfSCreateGroup to get the GrouDnld.

I

After a preempted run, this is the number of minutes to wait before attempting to rerun the job. If zero, the job will not be rerun after being preempted.

A preempted run is a job that did not run because the point product determined that it should not run due to point product specific operating rules.

Number of retry attempts to make if the job was preempted or was unable to start.

If zero, the job will not be rerun after preemption or start failure. This does not control reruns after a point product failure.

Number of failed (point product) runs to accept before the job is no longer rescheduled. This value is moved to "RecurFailureCount" upon job creation and every time the job runs successfully.

This field helps control runaway job rescheduling for jobs that never run correctly.

See "RecurFailureCountl, for more information.

WO 00/38033 PCT/US99/31024 0 0 ci 0 0 -i ci 1i 0 0 ci 29 RecurFailCount Instance Owner Cred value ConnectString JobOwner Object This function counts job failure for a given number of successive runs. It is decremented each time the job fails and reset to its original value (RecurFallureThreshold) each time a job is successful. If the value ever gets to zero, the job will not be rescheduled.

1 Database instance to run against.

Optional for database-related point products.

Optional user ID for point product use (For RDBMS product, this would be the owner.) Optional user password.

Encrypted/decrypted by API set and get.

Optional database connect string or name.

For future use related to security (Only the owner of the job can make changes.) Optional object name for point product use.

Table 7 lists internal functions that may be set by a user via the LES GUI.

I

WO 00/38033 PCT/US99/31024 30 TABLE 7 GUI Functions Function Description Jobld Unique job ID. Each unique ID is generated on the central control server. The IDs are unique within the nodes managed by that central server.

Use the function PtLESSubmitJob to set the JobID. The JoblD cannot be zero.

Next Start Indexed on local node. This is the next time, in GMT, to run this job. Time_t timestamp is set using the calendar and recurrence information the next time the job is started. If this job is to be run immediately.

TimeZone l=Use time zone of the node where the job will run.

2=Use time zone of the node from which the job is scheduled (local workstation's time zone).

This value from Window's time zone convention should be converted to whatever time zone convention LES uses. This function is used to convert the time to GMT. The API will retrieve the time zone's offset from GMT and adjust the time to GMT.

WSTimeZone NotificationEmail CompletionScript FailedScript Time zone of the workstation from which this job is scheduled. This is used by the scheduler to adjust the time if TimeZoneNode 2.

Email address where job completion notification will be sent.

Script to run upon successful completion.

Script to run upon failure.

\VO 00/38033 PCT/US99/31024 31 Description Optional job description.

Calendarld ID of the calendar to use.

Recurlnterval This Is used to test if the job fails for a given number of successive runs. It is decremented each time the job fails and is reset to its original value (RecurFaiiureThreshold) each time a job is successful. When the value equals zero, the job will not be rescheduled.

RecurIncrement How often to run the job, depending on recurrence interval; "Every x minutes, every x days, every x months, etc." EveryWeekday Days of Week (DayOfWeekFlags) DayCfMonthFlags This flag states to run every day week (RecurInterval Daily). If set and RecurInterval Daily, RecurIncrement is ignored.

of the this is Zero padded array. Flags to indicate which day or days of the week to run the job(0=Sunday, 6=Saturday) (RecurInterval weekly) Zero padded array. Flag to indicate which day or days of the month to run the job (1 to 31) (RecurInterval Monthly) WO 00/38033 PCT/US99/31024 +,1

O

O

0 0 o ci c-i ci ,1 ci 0 0 (1 32 LastDayOfMonth YearMonth TimeOfDayHour TimeOfDayMinute Start_Times If true, run on the last day of the month(s) (RecurInterval Monthly or Yearly). The job will run on the last day of the month regardless of which actual day the end of the month falls on (28, 29, or 31). Can be used with or without the DayOfMonthFlags. If set, this takes precedence over the DayOfMonthFlags the job flag for November 30 is off and this is on, the job will run on November Month to run (1-12) for yearly interval (RecurInterval Yearly) Hour of day (0-23) to run Minute of hour (0-59) to run Minutes after each hour flags. Zero-padded flags one for each minute of the hour.

This is used to emulate Autosys' run at x minutes after each hour behavior. Each flag, when on, signifies the job should run at that minute after EVERY hour after the first run which is determined by next start.

Table 8 provides example code for on how the GUI API 115 would be utilized by a point product. For example, allocating a job structure requires a point product to define a handle an array defining the days the job is to run (0=don't run, 1 run; ie, 0,0,0,0,0,1,0 means run on Friday), a userid and password, and a call to the PtLESJobAlloc() function.

WO 00/38033 PTU9/12 Cl33 o TABLE 8 0Step Sample Code Comment Allocate a P tL ESJo b Han dl11e j;run on Friday j ob int days_:o run (I structure Clchar userid Cl password jh=PtLESoobAlloc

C)

ClSet PtLEssetJoboescriptlon (3h, "ismith",, Retry eerparameters "TablespaceReorg", "Weekly run"); o min for anr, for the PtLSSsetJobCommand (jh, "reorgexer; hourev' job PtLESSetJobproa~nto (jh, "'Tsreorg',, week at PtLESSecJobInstanoeInfo "ora-73", userid,password, "payri world"); PtLESSetJobobjeot (jh, "payro'll); PtLESSetJobaetrcjh, 10, 6); FtLESSetJobTzNode (jh,PTLESTZLOCAL); PtLESSetsobRunRules (jh, PTLESRECURWEEKS, 1, 0 days_to run, NULL, NULL 0, 0, 0, 22, 0); Submit the PtLESSUbmitJcb Cjh, "dimultral,,jolo) ;node d~mulera job on two PtLESSubmitJob (J h,"dimspar 0 10 1 nrode dirr.sparc tnodes Clean up PtLESJobDealloc (jh) Calendaring-

API

The calendaring API is utilized internally by oth!er APIs and provides comprehensive calendaring funct-ions.

For example, main use of the Calendaring API is to allow users to input Calendaring information for a specific job run.

WO 00/38033 PCT/US99/31024 0

O

34 o Job Administration API o As discussed above, the job administration API 150 deletes, cancels, reruns, or copies existing jobs. In addition, it allows other APIs to determine job status, (Cq 5 and setup, cancel, or update a job logfile and retrieve e" job parameters. For example, one process within the Jcb C Administration API is the PtLESCancelJob utilized to stop currently running jobs. The PtLESCancelJob firs: o performs a check on a specified node to see if the job is running, and issues a stop command to the corresponding LES agent if the job is executing.

Job Data Management

API

Both the job and the job history are saved in the central repository and the local repository of the node where the job is to be run. Open Database Connectivity (ODBC) is used to select, insert, update, and/or delete each job or job history.

Point Product API The Point Product API allows the point product to communicate with the LES Agent. When you use this API, the LES job folder shows more accurate information about your job, the logfile viewer shows the job logfile, and the LES agent manages your point product job, making

LES

easier to use from a development standpoint.

The functions of this API encapsulate PEC messages to the LES agent to update the job status or logfile, or signal the end of a running job process. The PEC model is adhered to wherein the LES agent and the point product are Rtclients using broadcast communication.

PEC Initialization has been performed before calling any of the LES API functions. In addition, the Point Product should call TptTerm before exiting.

WO 00/38033 PCT/US99/31024 0

O

S- 35 O In one embodiment, SIGUSR2 is used as a CANCEL 0 notification signal on all UNIX platforms, and the LES agent ignores this signal by default. Because the child process inherits all signals, the CANCEL event is q r ignored unless the point product takes action. As a C result, conflicts may arise if a user application is Susing SIGUSR2 for other purposes (but allows for u:iliation of standard UNIX toolsets for operations) SThe following Environment variables are set before running any point product by the LES Agent: LESJOBID-The point product current job number LESRUNNUM-The Point product current run number LESGROUPID-The point product current strategy group number LES_RUNTIME-The schedule time(If the job was scheduled to run immediately.) LES_INSTANCE-The job instance LES_USERNAME-IThe job user name LES_USERROLE-The job user role Error messages are displayed whenever the agent detects an error. Table 9 lists the errors, consisting of a number and a message. Some of the messages use variables, which appear as signs in the message text as shown in the following example. These variables are replaced by actual values when the error message displays. For example, in this case the variable is replaced with the actual constraint and file name when the error message is displayed.

WO 00/38033 PCT/US99/31024 36 TABLE 9 PTLES_AGENT-0001 Agent is already running PTLESAGENT-0002 Cannot get the current node name PTLESAGENT-0003 Cannot open process table file PTLES_AGENT-0004 Cannot write to process table PTLES_AGENT-0005 Cannot read from process table PTLESAGENT-0006 Cannot allocate enough memory for the job handle PTLESAGENT-0007 Unable to free the job handle memory PTLES_AGENT-0008 Unable to find the job attached to this process id PTLESAGENT-0009 Unable to read job parameters PTLESAGENT-0010 Cannot allocate enough memory for the calendar handle PTLESAGENT-0013 Cannot create a PEC message (TipcMtCreat: failed %d) PTLESAGENT-0014 Unable to find and cancel the job with job id=%d and run number=%d PTLES AGENT-0015 The following is not a valid job, cannot cancel this job. (job id=%d, run number %d) PTLESAGENT-0016 Cannot cancel the following job (job id run number system error %d PTLES AGENT-0018 Cannot execute job process jobid %d and system error is: %d PTLES AGENT-0020 Cannot execute job process job_id %d, the user doesn't exist PTLES AGENT-0021 Cannot delete the job job_id %Id, error %d PTLES_AGENT-0023 PEC Callback create failed. (pec_error %d) i WO 00/38033 PCT'LISW31024 37 PTLES AGENT-0024 Cannot insert the job (jobid into the Local repository error °d PTLES AGENT-0025 Cannot insert a job history, job_id %d, error Agent initialization Checking the expired jobs: %s PTLES AGENT-0027 Cannot connect to the central repository- (ODBC-Error %d) PTLES AGENT-0028 Cannot connect to the local repository.

(ODBC-Error %d) PTLES AGENT-0029 Calendar insertion error(ODBC-Error %d) PTLES_AGENT-0030 Cannot update job (%d)(ODBC-Errcr %d) The agent is not running Table 10 provides example source code for a point product job template and may be considered an example use of the point product API 120. However, the code presented is not intended to be either a compilable or executable version of the present invention or a complete embodiment thereof, but merely an example representation of some of the features discussed herein.

TABLE Platinum Technology Inc Copyright 1997 Platinum technology Dimeric Lab All Rights Reserved.

PLATINUM Lightweight Enterprise Scheduler example program using the LES Point Product API referred to LES API Function source code.

Point Product job template and use of the Point Product API system includes WO 00/38033 PCT/US99/31024 32 #include <stdio.h> #include <stdlib.h> pec include files #include <rtworks/ipc.h> #include <ptm/ptm.h> LES_API include files #include "ptlessize.h" #include "ptles_ppapi.he edit keys here for example purposes the num of keys allowed is limited to #define MAXKEYS static char keys[MAX_KEYS] [KEY L+1]l "key e.g replace "key 1I "key 2-, "key 3", "key 4, with your key value #define MY_STATUS_RUNNING #define MY_STATUS_COMPLETE #define MY_STATUS ABORT #define MY_STATUS CANCELED FILE *fplog; If user want to cancel this job switch to the following function WO 00/38033 PCT/U JS9931024

O

S39

O

O void cancel this job( sig int sig; C fprintf(fplog,'Update Job Status to STATUSCANCELED if PtLESUpdateJobStacus MY STATUSCANCELED) =LESAPI FAIL

O

fprintf(fplog,uPtLESUpdateJobStatus: Error fprintf(fplog, Point Product example Job canceled by user request\n"); PtLESUpdateExitStatus(PPFAILED); TptTerm(); fclose (fplog); exit(l); example program demonstrates how a point product can use the LES_API to communicate with the LES Agent'/ int main( argc,argv int argc; char **argv; int rc; used for error return char pvalue [DATA_L+l]; allocation to hold value of a key-value pair*/ int plen; length of value int i; used as a counter int num_items; num of key-value pairs rec char *plat_home; char log_path[PATH_L+l]; char *1es jobid *job_id; char *les_groupid; WO 00138033 WO 0038033PCT/LS99/3 1024 0 char *les-runtime; O char *lea-run-number; #if defined(nEsuG) defined (W1N32) ci S Debug~reako; #endif o job_Id getenv('LEsJOBhIZ)}; plat_home getanv('PLATHOME); if( plat-home I- NULL jobid I- NULL Bprintf(log~ath, %a/lea/files/logdir/ppaogfile_%s.logm, plat-home, job_id); else strcpy(logpath, pplogfile.aogm); fplog fopen( log~path, *wO); if( Iplog

NULL)

fplog stderr; fprintf Pointproduct example Jlesjobid getenv('LES_-JOBID"); lea groupid getenv('LESGROUPID'); les-runtime getenv(OLESRUNTIME'); **LES V1.l.0 New Env.

lea-run-number -getenv(OLESRUNNUMBERO); 3 0 fprintf(fplog,mENV VARIABLE SETUP BY PTLESAGENT if( lee jobid I- NULL PtLESSetJobld(atoi (lea jobidfl; fprintf(fplog,* LEEJOBID %B \nu, lea jobid); W-0 00/38033 WO 0038033PCT/;US9913 1024 -41 if( les-groupid I- NULL fprintf (fplog,o LES_GROUPID %s \nm, les~aroupid); if( les-runtime

NULL)

fprintf(fplog,u,.... LESRUNTIME =s\n% if( lee-run-number I- NULL) fprintf(fplog,." LES RUNNUMBER las-run-number); les-runtine); %s \no 7* print our any arguments passed on the command line fprintf(fplog,mcoimmand line argument:\n) for im1;i argc;i++) fprintf (fplog, "Argument i rargv f* To use the LES API you always have to call the PEC rptlnit function*/ to make a connection to the RtServer in main program LES Vl.l.0 adapt. Add CANCEL Function PtLESSetJobCancel( &cancel this job fprintf (fplog, Hlpdate logfile name \no, log~path); Communicate the Point Products LogE ile to LES Agent if PtLESUpdateLogFileNama (logJpath)=-LEsAP:_FAIL fprintf(fplog. PtLESUpdateLogpileName: Error %d\nfl; Update the Point Product Status to Running WO 00/38033 PCT/US99/31024 42 use this function to update status fprintf(fplog,"Update Job Status to STATUS_RUNNING if PtLESUpdateJobStatus MYSTATUSRUNNING}==LESAPI

FAIL

fprintf(fplog tLESUpdatebtatus: Error d\ fprintf(fplog,"PtLESUpdateJobStatus: Error %d\nM); The following function call performs the initial communication with the LES AGENT Used to retrieve the point product parms that may be needed to be passed stored internally is the jobid and run number fflush(fplog); fprintf(fplog,-Get Point product Parameters: \na); rc=PtLESGetProductParms( 360 TIMEOUT if re LESAPIFAIL fprintf(fplog,-PtLESGetProductParms: Error %d\nR, PtLESGetErrno()); Another call to communicate the progress of a job to the LES AGENT fprintf(fplog,pp_example:Update the STATUS to MY_STATUSABORT %d

MYSTATUS_ABORT);

if PtLESUpdate Job S ta tu s W-O 00/38033PT/S931U2 PCT/US99/31024 -43 0 KY STATUS ABORT) ==LESAPI FAIL fprintf (fplog, PtLESUpdateJobStatus: Error %d]n C] Communicate exit status of point product pre defined to PPSUCCESS,PPFAILED, PPPREFMPTED o fprintf(fplog'upp_example .update the exit Status o %d \tPPWAILED); PtLESUpdatesxitStatus (PPFAILED)); PtLESflestroyParmsoL; TptTern fclose(fplog); exit edit the keys static char string defined before the main *to add point product keys return the no of key value pairs/ nun-items=PtLtSoetllunPans 01 if( nun itexns=LESAPI FAIL fprintf (Iplog, WPtLESGetNumParms: %d\nu, PtLESGatErrno 0); fprintf(fplog,*Wo. Parameters received:%d\n",nun items); nun-itemg=num-itemcflAxKEYS7'num-items:MAXKEYS; fprintf (fplog, wNo. Parameters received: %d\n num-i temls); retrieve an print the key value pairs w for(inO; i num-items; i+s+) 1* get the length of value in the key-value pair WO 00/38033 PCT/US99/31024

O

44- O plen- PtLESGetParameterLen(keys[i]); if( plen 0 Function used to retrieve the value given*/ a key for the parameter C4 if(PtLESGetParameter(keys ,pvalue, plen)== LESAPI FAIL) o fprintf(fplog, "PtLESGetParameter: %d\no, PtLESGetErrno{)); else print out the key-value pair fprintf(fplog, "\tParameter[%d] key%s Value=%s len=%d\n*, i,keys[i],pvalue,plen); print out the job Id and the Run Number received fprintf(fplog, Received Job Id: %ld\n',PtLESGetJobId()); fprint f fplog "Received Job Run Number:%ld\n",PtLESGetJobRunNum()); Another call to communicate the progress of a job to the LES AGENT fprintf (fplog,"Update Job Status to STATUSCOMPLETED \nw); if PtLESUpdateJobStatus (MYSTATUSCOMPLETE) ==LESAPI FAIL fprintf(fplog, 'PtLESUpdateJobStatus: Error Before terminating call PtLESExitStatus Communicate exit status of point product pre defined to PP SUCCESS,PP_FAILED,PP PREMPTED free the internal structures allocated by LESAPI*/ WO 00/38033 PCT!US99/31024 0 45 O fprintf(fplog,-Send END OFJOB with STATUS SUCCESS to o PTLES AGENT PtLESUpdateExitStatus(PP

SUCCESS);

PtLESDestroyParms (c 5 terminate connection to RTServer Cl TptTerm(); c fclose(fplog) return 0 eof of main Scheduling via POEMS The present invention has been implemented utilizing a GUI interface that includes either a window (POEMS Scheduling Service Job Scheduling Window, Figure or a property page (Job Scheduling tab property page, Figure 11). Other embodiments or arrangements of GUI windows or pages are also applicable, for example, using pull down selection menus instead of radio buttons, as just one example).

The GUI interface allows user input for scheduling specifications for a job to be submitted, including: Immediate job run.

Job run at a later time.

Starting time and date of the job.

Recurring run intervals for the job.

Create or select a scheduling calendar.

In addition, the GUI interface allows a user define or modify a notification script.

To start a job, a user enters a job description in the Job Description box, selects a run time as either immediate, a later time, or start date and time, and then Clicks on OK or Finish.

POEMS has several features for providing che WO 00/38033 PCT/US99/31024 46 appropriate start date and time. For example, as shown in Figure 12 a pull down calendar is provided for date selection. The pull down calendar includes advance and decrease buttons as follows: *Select the button tc advance the year vb ocn.

*Select the button to decrease the year by one.

*Select the button to advance to the next month.

*Select the button to move back to the previous month.

Since the POEMS scheduling service schedules jobs at remote nodes, a selection box is provided to use the local time of the remote node as the starting time (instead of the submitting node).

Recurring job runs may be set for a wide array of intervals, including: *None Used to schedule no recurring runs for the job.

'Minutes after each hour Used to schedule recurring runs for a job each hour at a specified number of minutes after the hour.

'Hours Used to schedule recurring runs for a job at specified hourly intervals.

*Days Used to schedule recurring runs for a job at intervals specified in days or every weekday.

*Weeks Used to schedule recurring runs for a job at intervals specified in weeks on the selected day or days.

"Months Used to schedule recurring runs for a job at intervals specified in months on the selected day or days.

*Years Used to schedule recurring runs for a job every year on the specified dates.

WO 00/38033 PCT/,'US99/31024 0 47

O

0 Figure 13 provides an example of a job scheduled to run at selected minutes after each hour (schedule the hourly recurring run interval). Figure 14 provides 5 an example of a job scheduled to run every hour C] (schedule the recurrinc run incerval in hours) Fiure Cq 15 provides an example of a job scheduled to run every Sday (schedule the recurring run interval in days).

o Figure 16 provides an example of a job scheeuled to run every week on Thursday with a specific start date (schedule the recurring run interval in weeks).

Figure 17 provides an example of a job scheduled to run every 25th day of the month (schedule recurring run interval in months). Figure 18 provides an example of a job scheduled to run once a year on June (schedule the recurring run interval in years) Notification Scripts As discussed above, the GUI interface provides the user an opportunity to define, modify, or select a notification script. Notification scripts are shell scripts containing actions used to automatically provide notification about jobs and job status information. Other functions may also be performed by the notification scripts, for example. The point product application executes the jobs notification scripts. The product application documentation should include detailed information on actions taken upon job completion or job failure.

Notification scripts may be constructed for tables, tablespaces, and indexes. When constructing a notification script for notification of a completed or failed job, variables of previously assigned values may be utilized.

WO 00/38033 PCT/US99/31024 48 The values for these variables are the same ones defined for a particular job in the defaul notification script files. A user may define the values for the variables in these files, and then use the variables in the notification script as arguments.

Using these variables in a notification scripallows the same script to be utilized for various jobs, since different values may be assigned to the variables for each job. You can use variables in a job completion script or job failure script as determined by the product application.

Table 11 provides a listing of established notification script variables, including a description of each variable.

TABLE 11 This Script Variable Represents This Value %C(JOB_ID) Job identification number %C(GROUP ID) Group identification number %C(HOSTNAME) Name of the host on which the completed or failed job ran %C(INSTANCE) Instance name on which the completed or failed job ran %C(USERNAME) Name of the user %C(CONNECT_STR) Connect string %C(JOB_OWNER) Owner of object whose job completed or failed %C(OBJECT) Name of object whose job completed or failed %C(PRODUCT CODE) Product Code %C(PRODUCT VERSION) Version number of the product %C(DESCRIPTION) Job description 1 WO 00/38033 PCT/US99/31024 0

O

49 0 0N %C(EMAIL_ADDR) Notification routing string (the value Cl entered in the Notification email address field) %C(TYPE) Job type code, representing the type cf Sjob that completed or failed (CPLATHOME) Install directory %C (USERROLE) The role assigned to the user e.g., 0 administrator A Notification Scripts button on the GUI interface accesses a Notification Scripts window (see Figure 19) The Notification Scripts window allows a user to: *Modify the text to use for confirmation messaces indicating the completion or failure of a job.

*Write a customized script using the variables specified in this window.

*Specify the path for the notification email address used to inform you of the job's completion or failure.

In each of the Job completion script box and Job failure script box, a user may enter any of the predefined script variables to modify default scrip:, or simply type in a full path location and filename for a notification script.

A full path location for a user or notification email box address is typed into the Notification email address box. This method may be utilized to assign notification routing addresses for a pager, email, or phone number to deliver the information about the scheduled jobs to any location.

WO 00/38033 PCT/US99i31024 0 0 S50

O

O Schedulin a Calendar 0 A Calendar Selection window, see Figure accessible from the GUI interface, provides a list of the predefined calendars that may be used with the current job. Using the options in this window, a user C can: C] 'Create a new calendar.

o 'Edit an existing calendar.

0 'Select a previously defined calendar.

'Delete a calendar To create a calendar, a user selects the Create button in the Calendar selection window, which invokes the display of a Create Calendar window (see Figure 21). The user fills out the ID (a name for the calendar) and Description fields.

The user may select the following buttons to select the month and year of a date on which the job is to run: 'Select the button to advance the year by one.

'Select the button to decrease the year by one.

*Select the button to advance to the next month.

'Select the button to move back to the previous month.

By clicking a day of the month, a red border marks the date on the calendar and the complete date displays in the Selected dates field. Multiple days may be selected.

Existing calendars may also be modified by selecting a calendar first by clicking on its ID or Description in the available Calendars list in the Calendar Selection window and then clicking on the Edit button (see Figure 22, for example) Once the calendar is displayed, additional job run dates may be added by WO 00/38033 PCT,/US99/31024 0 S- 51

O

o selecting a date on the calendar. A user utilizes: *Select the button to advance the year by one; *Select the button to decrease the year by one; 'Select the button to advance to the next month; *Select the button to move back to the previous C] month; and o then the user clicks on the day of the month. A red C 1 border marks the date on the calendar and the compleze date displays in the Selected dates field. The above process is repeated until all the dates for runnirg the job are displayed in the Selected dates field.

Dates may be deleted from the Selected dates lisby clicking on calendar numbers marked with red borders to deselect. The border on the date disappears when deselected and the date no longer displays in the Selected dates field.

The OK button saves the calendar and exits the window. The edited calendar is then available in the Calendar Selection window. Clicking the Cancel button exits the window without saving changes to the calendar.

A calendar may be selected by clicking on the Select button of the GUI interface (Job Scheduling window/property page), which displays the Calendar Selection window (Figure 20, for example). A calendar is highlighted by clicking on the calendar ID or description in the Available Calendars list, and then clicking on the Select button (or double clickina on the ID or description). The calendar selection window then closes and the selected calendar ID displays in the Calendar ID field on the Job Scheduling window or property page.

A calendar my be deleted using the above procedure WO 00/38033 PCTiUS99/31024 0 0 52 0 o by clicking the Delete button instead of the Select C button from the Calendar Selection window.

C] Strateyv Scheduling A Strategy Scheduling window is provided to view, C] create, modify, or delete schedules for a stratec-v.

The schedule strategy window is invoked from a point Sproduct, as shown in Figure 23.

C] A Create button is provided for creating a new schedule for a current strategy, which invokes the Job Scheduling window/property page (see Figure 24) The Job Scheduling window/property page is then filled out for the new schedule.

An Edit button is provided to modify an existing schedule for the current strategy, which invokes the Job Scheduling window/property page available for editing a selected schedule.

A delete button is provided to delete a schedule for the current strategy. A user first highlights a schedule to be deleted and then clicks on delete. A Delete Schedule window prompting for a clean up script is then displayed (see Figure 25, for example) The user may either delete the schedule without running a cleanup script by deleting the text (if any; in the Name of a shell script or other process to run to clean up associated files field, or delete the schedule and run a cleanup script by typing a path and full name of the script. The user invokes the action by clicking OK or returning to the Strategy Scheduling window without deleting by clicking Cancel.

Job Management Services The present invention includes multiple job management services that are presented in an easy to WO 00/38033 PCT/US99/31024 0 c, 0 53 o use and intuitive format. The present invention c' utilizes a Jobs resource object to allow a Director program to identify and track various job processes.

After inserting a Jobs resource object into the Director, a user may perform the following jcb management tasks: 'View information in the following formats: S*Columns Located on the right side of the C] Director window.

'Property pages Located on the right side of the Director window.

*Logfile Supplied Log File Viewer *Delete jobs 'Rerun jobs 'Cancel job runs *Monitor the progress of job runs A Jobs resource object provides a bookmark for finding information to display, and locating a job to be acted on when utilizing job management processes (deleting or monitoring a job, for example) An Insert Object window (see Figure 26), invoked by the user, identifies objects to insert (POEMS provides a number of objects that may be inserted, a partial listing is provided in Figure 26). The user selects an object (POEMS Scheduling Service, in this example), and presses OK. In response, a properties page is displayed (see Figure 27), which includes General (Figure 28), Representation (Figure 29), Subscription, Indicators, Job Repository (Figure 30), and Event Correlation tabs.

The Properties page includes a label and description field. The user enters a label and a description (description is optional).

WO 00/38033 PCT/US99/31024 0 0 c, -q- 0 54-

O

O The Representation tab allows the user to select, CA in an Intellicon View, an icon for representing the job, and optionally, a background and background color.

Alternatively, the user may select an icon in Explorer 5 view.

C]

cq The Intellicon view is show in Figure 29, and includes a list of Intellicon icons (provided by POEMS files, or alternatively provided by the user), and a C( display window to view a selected icon (graphic) An example of the Explorer view is illustrated in Figure which is an Explorer view displaying Provision (point product, or application) Jobs (or LES jobs) as the label for a Jobs resource object.

Selecting the Job repository tab displays a window for defining a timezone for display of times related to jobs. A down arrow pull down menu bar provides the user with selections of any timezone. This selection does not affect the time or timezone selected for running a job as specified in the Job Scheduling property page. In one embodiment, daylight savings time may be compensated for by providing an automatic compensation selector that invokes a program to adjust for daylight savings time; alternatively, the user may select a timezone just east of the desired timezone.

Using Job Resource Obiects The Job Resource objects defined by the above processes enables a user to locate and view both jobs and runs. The user can locate each job in the Explorer view. Each time a job executes, it creates a new run, which then displays in a run folder for the job.

Each Jobs resource object contains a hierarchy of folders sorted into the following categories: WO 00/38033 PCT/US99/31024

O

o 55 o *All Jobs Lists all jobs regardless of current C] status and also lists all runs by status.

*Jobs By Group Lists all jobs arranged by group C Ids.

1 5 'Jobs By Node Lists all jobs arranged by the C] node on which the job runs.

C]

-Jobs By Product Lists all jobs arranged by the Sproduct used for the job.

C] 'Jobs By Type Lists all jobs arranged by type.

'Jobs By User Lists all jobs arranged by the user who scheduled the job.

However, a point product may not allow a user to assign a job group or type. The hierarchy of folders is illustrated in Figure 32.

The All Jobs Folder (see Figure 33) numerically lists all of the jobs instead of categorizing them by group, product or type. The All Jobs Folder lists jobs in folders, including: *All Jobs Any Status Lists all jobs regardless of status along with associated job history (each run of the job).

*All Runs By Status Arranges all of the runs of jobs into the following folders according to their current status: *Completed Runs 'Failed Runs 'Not Started Runs 'Preempted Runs 'Running Runs 'Stopped Runs *Held Jobs Lists all of the jobs that are held and can be scheduled later.

WO 00/38033 PCT/US99/3 1024 0 56 o .Scheduled Jobs Lists all of the jobs that are scheduled to run.

C, A user can display jobs according to the groups assigned to them when they were originally scheduled by C] using the Jobs Group Folder. The specific producapplication (used to run the job) assigns a group, but o may not use this classification.

-c Jobs may be displayed according to the node on which they ran by using the Jobs By Node Folder.

Jobs may be displayed according to the specific product that ran each job by using the Jobs By Product Folder. This is helpful if using multiple ProVision products to schedule and run jobs with the POEMS Scheduling Service (helpful because it allows grouping of jobs by function or application, for example).

Jobs may be displayed by job type assigned to them when the jobs were originally scheduled by using the Jobs by Type Folder. The user's specific product application (used to run the job) assigns the job type.

However, products do not necessarily use this classification.

Jobs may be displayed according to the user who scheduled the job by using the Jobs By User Folder.

Each of the Jobs By User, Jobs By Type, Jobs 3By Product, Jobs By Node and Jobs By Group Folders contain the same folder hierarchy for each user as that for :he All Jobs Folder described above.

Specific jobs may be located by expanding folders using the signs in the Jobs resource object hierarchy. The resulting display includes Job ID, Job Icon, and a Job description.

In addition, a sign of a specific job reveals its run history (See Figure 35, for example).

WO 00/38033 PCT/US99/31024 c,

O

0 57-

O

The run history is stored in several folders. The CA All Runs folder contains every run regardless of its status. The remaining six folders contain only :he F' runs that possess the status that applies to Tha: 5 folder (Completed, Failed, Not Started, Preemted, q Running and Stopped Runs).

T_ Specific runs may be located in the All Runs g folder and according to its status as illustrated in C( Figure 34. In Figure 34, Run #1 of Job 883 is stopped; therefore, it displays in the Stopped Runs folder as well as in the All Runs folder.

Each run listing includes: *Run icon *Run number *Run start date and time Data can be viewed in column format for all jobs, runs, groups, nodes, products, types and users stored in a job, runs, groups, nodes, products, or types folder. The data displays in columns on the right side of the Director window when using the Explorer view.

As shown in Figure 36, some of the data is shown that is available for the jobs in the All Jobs Any Status folder under a specific node in the Jobs By Node folder.

A user can select any of the following folders to view data for their contents in column format (note that some folders, such as All Jobs, do not provide data), including: *Jobs By Group *Jobs By Node *Jobs By Product *Jobs By Type *Jobs By User 'All Jobs Any Status WO 00/38033 PCT/US99/31024 0 ci 0 58 0 o *All Runs By Status C 'Held Jobs *Scheduled Jobs SAll Runs 5 Completed Runs Ci 'Failed Runs *Not Started Runs o 'Preempted Runs C -Running Runs 'Stopped Runs Viewing Job Group Data The following data columns display when a user clicks on the text of the Jobs By Group folder in a Jobs resource object: *Group ID Strategy group ID number.

*Strategy ID Strategy ID number of the strategy defined in the Common Strategy Services to which the group belongs.

'Description Description of the group that the user can enter when the group is created.

*Product Code Code for the product used to create the jobs that can be viewed in this folder.

*Product Version Version number of the product used to create the jobs that can be viewed in this folder.

'Jobs_PerRun Number of POEMS Scheduling Service jobs submitted by the product for this strategy group.

*Created Date and time the strategy group jobs were generated.

WO 00/38033 PCT/US99/31024 0 0 59 Viewing Node Data The following data column displays when the user clicks on the text of the Jobs By Node folder in a Jobs resource object: Node List of the nodes under which jobs are C- located.

O Viewing Product Data (Ci The following data columns display when the user clicks on the test of the Jobs By Product folder in a Jobs resource object: 'Product Code Code for the product used to create the jobs that can be viewed in this folder.

'Product Version Version number of the product used to create the jobs that can be viewed in this folder.

Viewing Data Type The following data column displays when the user clicks on the text of the Jobs By Type folder in a Jobs resource object: *Type List of job types. The user can locate jobs under their job type, which is assigned in the product used to run jobs.

Viewing User Data The following data column displays when the user clicks on the text of the Jobs By User folder in a Jobs resource object: 'User List of user names. The user can locate jobs listed under user names.

WO 00/38033 PCT1tJS99/31024 0 ci o 60 0 o Viewing Job Data c The following data columns display when the user clicks on the text of an All Jobs Any Status, Held C. Jobs, or Scheduled Jobs folder in a Jobs resource object: C "'Job ID Unique Job ID number.

SDescription Optional description of the job.

S*Job Group Strategy group ID number for the

C

group to which the job belongs.

'Access Mode If this column displays L, the job is locked and you cannot rerun the job until the job is unlocked. A job is locked and unlocked by the point product running the job. If a job is not locked this column remains blank.

'Product Product that created the job.

'Type Job type (product-dependent).

*Node The machine on which the job will run.

'Whosetz Time zone to use for schedulino -he job.

'When Deployed Time that the job is written into the job table.

'Next Start This field may contain one of the following values or text messages: 'The next time the job is set to run.

*Run immediately.

'Expired.

*Schedule later (held) *Run Count Total number of times the job was run.

Viewing Run Data (Job History) The following data columns display when the user clicks on the All Runs By Status folder and any folder containing runs: WO 00/38033 PCT/US99/31024 61 0 0 61 'Job ID Unique Job ID number for this run.

Ci 'Run Number Unique run number (the number assigned to each recurring run of the job) 'Group ID Strategy group ID number of the arou where the job belongs.

C Product PLATINUM ProVision product (or o:her i product) that created the job.

O Type Job type (product-dependent).

Cq 'Time Zone Time zone used for job runs.

*Scheduled Start Time Time the job is scheduled to start.

*Start Status Status of how the job started.

This status is set by the POEMS Scheduling Service and the column may contain the following values: Job started successfully.

Job did not run because the startinc time passed while the POEMS Scheduling Service was down and the job was not recurring.

Could not execute due to OS status.

Fork failed due to OS status (insufficient system resources).

Invalid user.

*Actual Start Time Time when this job run actually started.

*End Time Date and time when the run completed.

*OS Status Provided by the operating system when job process could not be created (see your operating system documentation for information).

'Complt Status Completion status of the run assigned by the point product that ran the job (see your product-specific documentation for information).

'Failed This column may contain the following values: 1 WO 00/38033 PCT, 31(124 0 0 S- 62 S0: Job was successful.

C *I1: Job failed for any reason.

'Preempted This column may contain the following values: 5 Job did not run because it was disallowed due to your specific product's operaing OC rules.

O Job was not preempted.

C *Stopped This column may contain the following values: Job process finished without notifying the agent and the status is not known.

Job was not stopped.

*Logfile Name and location of the logfile for this run.

Viewing Data in Property Pa-e Format Data about specific objects (jobs, runs, groups, products, or types) may be viewed within a Jobs resource object. The data displays in a property page format with one or more tabs on the right side of the Director (Explorer view) window. Clicking on the text of the object displays the data in property page format on the right side of the Director, as illustrated in Figure 37, accessing the property page data.

The General property page contains the following fields: *Job ID Unique Job ID number.

*Description Optional description of the job.

'Group ID Strategy group ID number of the group to which the job belongs.

*Target Node The machine where the job will run.

WO 00/38033 PCT/US99/31024

O

0 ci 0 63

O

'Deployed to Target Node Time when the job is (C written into the job table on the node where it will run.

-e *Job Type Job type (product-dependent).

5 'Job Owner Owner of the object on which the job is performed.

CA *Run Count Number of times the job was run.

o *Run State This column may contain the foilowing

O

eC values: Waiting for the next run.

Currently running.

*Scheduling Time Zone Time zone of the workstation where this job was scheduled.

*Retry Interval (Minutes) This column may contain the following values: *Number of minutes to wait before attempting to rerun the job after a preempted run Or job will not rerun after it is preempted.

'Retry Count This column may contain the following values: 'Number of retry attempts made on the job when it is preempted or is otherwise unable to start.

Or the job will not be rerun.

*Recur Fail Threshold Number of failed product runs accepted before the rescheduling of the job is stopped.

-Recur Fail Count Used to test if the job has failed for a given number of successive runs.

(The number is decreased with each failure and reset to the original value with each successful run; if it is 0, the job is not rescheduled.) WO 00/38033 PCT-/US99/31024 0 0 64 O 'Access Mode If Locked appears, the user cannot (q rerun the job until the job is unlocked. A job s locked and unlocked by the point product runninc the job. If a job is not locked, this field remains blank, C Other Property pages are displayed by clicking on a corresponding tab. The Command property page (see 0 Figure 38) contains the following fields: 'Command Line Command line to execute.

'E-Mail Completion Notification to Email address used to send job completion or failure notification.

*Completion Script Job completion script.

'Failure Script Job failed script.

'Product Code Code for the ProVision product that created the job.

The Databases property page (see Figure 39) contains the following fields: 'Database Connect String Optional database connect string or name used.

'Database Instance Name of the instance on which the completed or failed job ran.

*Object Name Name of object whose job completed or failed.

*User ID Database user ID.

'User Password Encrypted database user password.

Note: The information on this property page is set by the point product that runs the job.

The Job Scheduling property page (see Figure displays the following job scheduling selections made in the product that ran the job: WO 00/38033 PCT/US99/31024

O

c, 0 65

O

'Start date and time of the job.

CA *Time zone where the job was scheduled.

Note: If the Use time zone of this workstation button is selected, it refers to the workstation where the jcc was scheduled.

C( 'Interval at which the job reruns.

C* 'Calendar ID of stored calendar used tc run the job. This field displays a Calendar ID only if a C] stored calendar was used.

The parameters tab property page (see Figure 41) contains the following field: 'Job Parameters Lists the jobs parameter names and values. (The point product determines the contents of this field.) The General tab property page (see Figure 42; contains the following fields: *Job ID Unique Job ID number for this run.

'Run Number Unique run number (the number assigned to each recurring run of the job).

*Time Zone Time zone where the job runs.

'Scheduled Start Time Time when the job is scheduled to start.

'Actual Start Time Time when this run of the job actually started.

'Start Status Status of how the job started (set by the POEMS Scheduling Service) *Started successfully.

'Expired Job did not run because the start time passed while the POEMS Scheduling Service was down and the job was not recurring.

WO 00/38033 PCT/US99/31024 0

O

0 66 0 'Agent down The POEMS Scheduling Service C( agent is down.

'Fork failed Process could not run on agenT machine due to insufficient system resources.

5 *End Time Date and time the run ccmoleted.

C]

C- *Operating System Status Status of the process

C]

provided by the operating system.

S'Run Status One of the following radio buttons C] is selected, indicating the current status of the run: 'Completed The run is finished.

'Not Started The run has not started yet.

*Running The run is currently in progress.

*Stopped The process has finished without notifying the agent and the status is not known.

*Preempted The run was disallowed due to point product operating rules.

*Failed The run failed due to an unspecified reason.

*Completion Status Code Completion status of the run assigned by the point product that ran the job (see product-specific documentation for information).

*Logfile Name of the logfile for the job.

To access the Run Stats tab property page, the user clicks on the Run Stats tab to view data on the Run Stats property page (see Figure 43).

The Statistics field displays product-specific information about this run of the job. This field only displays data if data is provided by the Provision, Point or other product.

WO 00/38033 PCT/US9!31024I ci 0 0 0 67

O

O Viewing Job Group Data Ci To Access job group data: *Click on the text of a group in a Jobs resource object, the General tab property page, as shown in Figure 44, is displayed.

C- The General tab property page (for a group) contains the following fields: S'Group ID Strategy group ID number.

C *Description Optional description of the group.

*Strategy ID Strategy ID number of the strategy defined in the Common Services Strategy where the group belongs.

*Product Code code (three-letter Platinum code, in one embodiment) for the Provision product used to create the jobs that can be viewed in this folder.

*Product Version Version number of the ProVision product used to create the jobs, which can be viewed in this folder.

*Jobs Per Run Number of POEMS Scheduling Service jobs submitted by the product for this strategy group.

'When Created Date and time the strategy group was generated.

Viewing Log File Data Log files are generated by various Provision products, point products, application products, etc (a class of software products that include links to the APIs and related programming as discussed herein) when jobs are run. If created, the user can view the log files created by products using the POEMS Scheduling Service through a Jobs resource object. The user may launch a log file viewer.

WO 00/38033 PCT/US99/31024

O

0 0- 68

O

STo determine if a log file exists for a given job, C the user first clicks on the text or the icon of an All Runs By Status folder to display the data in column format on the right side of the Director window. Then, scroll left to view the LOGFILE column using tne c horizontal scroll bar or the right arrow (see Figure 45, for example). If a log file for a run is available, its location displays in the row for that C] run in the LOGFILE column.

To view a log file for a run, the user finds the run of the job in a Jobs resource object. Then, right clicks on the run to display a popup menu (see Figure 46, for example). Finally, selecting View Log File from the popup menu brings up the display shown in Figure 47.

Deleting Jobs When the user no longer requires a job history, it may be deleted using the Jobs Administrator window.

Either multiple jobs or single jobs may be deleted.

Deleting Multiple Jobs To delete multiple jobs, the Jobs Administrator window is accessed at the level of the folder that contains the jobs to delete in a Jobs resource object.

Then, a right-click on the icon or text of the folder to display the popup menu illustrated in Figure 48.

Selecting Delete from the popup menu displays the Delete Jobs window (see Figure 49). This window displays the Job ID and node name of all of the jobs selected in the Submit Jobs/Runs field.

All of the jobs in a folder are initially selected and highlighted. If the user clicks on a job, it is deselected and will not be deleted. If all jobs are WO 00/38033 PCT/LIS99/31024 0 0 69 0 deselected, the Submit button becomes inactive.

0- Deselect any jobs by either: 'Click on the jobs one at a time (as shown in Figure 49); Or Use the Select None button to deselect all the jobs, and then click on the jobs to delete; 0. Or 'Click on the jobs to delete one at a time to initially deselect them, and then use Invert button to reverse your selection. (This will reselect the jobs you deselected and deselec all others.) Note: The Select All button may be used to reselect all the jobs.

Clicking on the Submit button deletes the selected jobs. The window expands to provide a dynamic display of the status and results of the action (see Figure If the deletion of a job completes successfully, the job ID and node name for each job display in the Succeeded field. If the deletion of a job fails, the job ID, node name, and an error code for that job display in the Failed field.

A running total of submitted, succeeded, pending and failed deletions displays in the small Submitted, Succeeded, Pending, and Failed fields. Alternatively, a user may click on the Close button while the jobs are deleted without waiting for all of the results tc display in the expanded window; or wait until all of the results (successful or failed) display in the Succeeded or Failed fields in the expanded window. The Close button then changes to the Done button, which vou WO 00/38033 PCT/US99/31024 070 0 0 select to close the Jobs Administrator window.

Deleting Single Jobs To delete single jobs, the Jobs Administrator Cq 5 window is accessed at the level of the job to delete C] a Jobs resource object. The user then locates the .ob (c to delete and right-clicks on the icon or text of the job to display a popup menu (see Figure 51) and selects O Delete.

The Delete Jobs window then displays (see Figure 52). This window shows the Job ID and node name of the selected job. If the user clicks on the job, it is deselected and the Submit button becomes inactive.

Clicking the Submit button deletes the job, and the window expands to provide a dynamic display of the status and results of the action (see Figure 53,.

If the deletion of the job completes successfully, the job ID and node name for the job display in the Succeeded filed. If the deletion of the job fails, the job ID, node name, and an error code for the job display in the Failed field.

A running total of submitted, succeeded, pending, and failed deletions displays in the small Submitted, Succeeded, Pending, and Failed fields.

Rerunning Jobs The Jobs Administrator window may be utilized to rerun completed jobs using the same parameters. Either multiple jobs or single jobs may be rerun. When a :ob is rerun, a new run number is assigned. In one embodiment, locked jobs are prevented from being rerun via these procedures.

WO 00/38033 PCT/US 99/31024 0 71 0 o Rerunning Multiple Jobs c 1 To rerun multiple jobs, the Jobs Administrator window is accessed at the level of the folder C] containing the jobs to rerun in a Jobs resource object.

The folder containing the jobs to rerun is located, and C] a right-click on the icon or text of the folder

C]

displays a popup menu (see Figure 54). Selecting Rerun o from the popup menu displays the Rerun Jobs window (see (C Figure 55). This window displays the Job ID and node name of all of the jobs selected in the Submit Jobs/Runs field and provides options for rerunning them.

Similar to the Delete Jobs procedures discussed above, all of the jobs in a folder are initially selected. If the user clicks on a job, it is deselected and will not be rerun. If the user deselects all of the jobs, the Submit button becomes inactive.

A user may utilize one of the following options in the Start Date Time specification area by: *Clicking on the Run Immediately button to rerun the jobs immediately; Or 'Clicking on the Schedule later button to cancel the previously scheduled next starting time for the jobs and hold the jobs in the Held Jobs folder.

Note: Jobs may be rerun either immediately or at a specified time using the Job Administrator window (see Rerunning Multiple Jobs discussed above).

WO 00/38033 PCT/US99/31024 S72

O

o Or 'Clicking on the Run at button and type the date in the Job Start Date field, or clicking on the C] down arrow to select the date from a calendar, then typing the time in the Job Start T"mer.e ld, C] or using the up/down arrows to scroll to the desired time.

"C Note: In one embodiment the user must set the date and time in these fields using the time zone originally used to run the jobs. The job times in the Jobs resource object may not display in the time zone where the job originally ran.

The selected jobs are then rerun at the specified date and time. Clicking on the Submit button reruns the selected jobs, and the window expands to provide a dynamic display of the status and results of the action (see Figure 56).

The selected jobs are then rerun at the specified date and time. Clicking on the Submit button reruns the selected jobs, and the window expands to provide a dynamic display of the status and results of the action (see Figure 56). If a job is submitted successfully, the job ID and node name for each job display in the Succeeded field. If a job fails to submit, the job ID, node name, and an error code for that job display in the Failed field.

A running total of submitted, succeeded, pending, and failed reruns displays in the Submitted, Succeeded, Pending, and Failed fields.

WO 00/38033 PCT/US99/31024 S- 73

O

O Rerunning Single Jobs C To rerun single jobs, the Jobs Admrinistrator window is accessed at the level of the job to rerun.

After locating the job to rerun in the Jobs folder, the icon or text of the job is right-clicked to display a

C]

C] menu (see Figure 57), and Rerun is selected, displaying the Rerun Jobs window (see Figure 58). This window o displays the Job ID and node name of the selected job C] and provides options for rerunning it. Clicking the job deselects it and the Submit button becomes inactive. Similar start date time options, as discussed above, are also available. Clicking on the Submit button reruns the job.

As with the other rerun options, the window expands to provide a dynamic display of the status and results of the submission.

Canceling Runs The user can use the Jobs Administrator to cancel a running job. The user can cancel multiple runs or single runs of a job. However, some products may temporarily disable this capability during a critical stage of a job, which does not allow recovery, makes recovery difficult, or requires that the product specific procedures be followed for recovery.

Cancelina Multiple Runs To cancel multiple runs, the Jobs Administrator is accessed at the level of the folder containing the runs to cancel in a Jobs resource object. After locating a Running Runs folder containing the runs to cancel, a right-click on the icon or text of the folder displays a popup menu (see Figure 59).

By selecting Cancel from the popup menu, a Cancel WO 00/38033 PCT!US993102 0 74 0 O Runs window is displayed (see Figure 60) This window C displays the Job ID, run number, and node name of all of the runs selected in the Submit Jobs/Runs field (see Fig. 57, for example).

Similar to the Delete Job procedures, all of the

C<

C] runs in a folder are selected initially. By clickinC on a run, it is deselected and will not be canceled.

o If all of the runs are deselected, the Submit button C' becomes inactive, etc. The Submit button cancels the selected runs, and the window expands to provide a dynamic display of the status and results of the action.

Monitoring the Progress of Jobs A Progress Monitor is provided to view information about a current phase and overall progress of any job run by products using the POEMS Scheduling Service, Individual products publish events regarding the different phases of their jobs. The Progress Monitor subscribes to these events and uses the resulting data to provide the user with information on the progress of the job.

The specific progress identifiers and the job phases that may be monitored are dependent on the individual design of the product using the POEMS Scheduling Service.

Accessing the Progress Monitor The Progress Monitor may be accessed at the individual run level in the Jobs Repository Resource Object.

To monitor the progress of a job, a user first finds the run of the job and right-clicks on the run to display a popup menu (see Figure 61). Selecting WO 00/38033 PCT/US99/31024 S- o Progress Monitor from the popup menu displays the C Progress Monitor (see Figure 62) The Progress Monitor is configured to display the following information: 'Name of the current phase of the job.

'Job ID and run number.

C] 'Name and/or number of the current phase or the job.

'Current phase number and the time remaining in (C the current phase.

'Graphical display of the progress of the current phase of the job and the percentage completed of the current phase.

*Graphical display of the overall progress of the job (including all phases) and the percentage completed for the entire job.

In one embodiment, the contents of the main text field vary according to the design of the product that scheduled the job.

The name of the current phase of the job can be viewed from the title bar of the Progress Monitor window. The progress of the job's current phase is determined by viewing the main text field in the Progress Monitor window, or viewing the Current Phase Progress field in the Progress Monitor window. Black bars are used to graphically display the phase progress. This field also provides the percentage completed of the job's current phase.

The progress for the entire job is viewed using the Overall Progress field in the Progress Monitor window. Again, black bars are used to graphically display the overall job's progress, and the percentage completed for the entire job.

WO 00/38033 PCT/US99/31 024 -76 o Troubleshooting Tips C] The present invention includes a number of troubleshooting techniques. A number of possibilities r q may cause a failure to delete and rerun jobs, or cancel specific runs of a job.

C] If a submission fails:

C]

SCommunications to the local and central o repositories may not be working. Verify that you c] can communicate with both of the repositories, and then submit the request again.

*The user may have attempted to cancel a run during a critical phase of job execution. The specific Provision product being used may prevent the user from using the cancel feature to avoid a potential conflict.

*The POEMS Scheduling Service agent (ptlesag.exe) may not be running on the node/machine where the object is located. Verify that a scheduling service agent is running on that machine, and then submit the request again.

To verify that a Scheduling Services agent is running on a node, a user selects Tools 0 Monitor Service Managers from the menu bar to display the Director Service Manager Monitor window (see Figure 63). The user may click on the plus sign or double-click on the text of Service Managers to display the list of service managers on different nodes.

Clicking on the text of the service manager for a node brings up the right side of the Service Manager Monitor with information for the programs residing on that node (see Figure 64).

The present invention has been described in reference to the POEMS Scheduling service utilized by WO 00/38033 PCT/US99,31 024 0 0- 77 0 O the Platinum suite of database products. However, the C] teachings of the present invention may be applied to any individual or suite of computer or other products to provide similar services. Therefore, the present invention is not limited to a specific line of products

C]

C] or type of products (point products, Provision

C]

products, or databases, for example) and apply to any Scomputer processes or applications in general.

C] The present invention may be conveniently implemented using a conventional general purpose or a specialized digital computer or microprocessor programmed according to the teachings of the present disclosure, as will be apparent to those skilled in the computer art.

Appropriate software coding can readily be prepared by skilled programmers based on the teachings of the present disclosure, as will be apparent to those skilled in the software art. The invention may also be implemented by the preparation of application specific integrated circuits or by interconnecting an appropriate network of conventional component circuits, as will be readily apparent to those skilled in the art.

The present invention includes a computer program product which is a storage medium (media) having instructions stored thereon/in which can be used to program a computer to perform any of the processes of the present invention. The storage medium can include, but is not limited to, any type of disk including floppy disks, optical discs, DVD, CD-ROMs, microdrive, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of media or device suitable for storing

I

WO 00/38033 PCT/US99/31024 0 0 ci -q- 0 78

O

0 instructions and/or data.

C1 Stored on any one of the computer readable medium (media), the present invention includes software for controlling both the hardware of the general S 5 purpose/specialized computer or microprocessor, and for Cq enabling the computer or microprocessor to interact with a human user or other mechanism utilizing the Sresults of the present invention. Such software may CRl include, but is not limited to, device drivers, operating systems, and user applications. Ultimately, such computer readable media further includes software for performing the present invention, as described above.

Included in the programming (software) of the general/specialized computer or microprocessor are software modules for implementing the teachings of the present invention, including, but not limited to, application programming interfaces for Graphical User Interfaces, Job Scheduling, Job Data Management, Job Administration, a command line interface, calendaring functions,and communications across a network. The software modules also include job scheduling agents operating on individual nodes on computer platforms in the network, and modules for the display, storage, or communication of results according to the processes of the present invention.

Each of the above-described APIs are compiled and linked into specific point products, ProVision products, or other applications utilizing those APIs to perform the present invention. In addition, configuration files are maintained for setup and operation of the invention and repositories are maintained by the invention.

Obviously, numerous modifications and variations WO 00/38033 PCT/U S99/31024

O

o 79 o of the present invention are possible in light of the

C

1 above teachings. It is therefore to be understood that within the scope of the appended claims, the invention may be practiced otherwise than as specifically 5 described herein.

0 0

(N

Claims (35)

1. A job scheduling device for scheduling jobs to C run on at least one node of at least one computing C] platform, comprising: o an enterprise scheduling agent installed cn each C] node and configured to launch execution of jobs submitted to the agent; a presentation layer configured to accept and validate parameters identifying at least one job to be submitted for execution on at least one of said nodes; and a job scheduler configured to allocate at least one job based on said parameters and submit the allocated jobs to at least one enterprise scheduling agent.

2. The job scheduling device according to Claim 1, further comprising: a job data management device configured to maintain job data and job histories of jobs submitted to each enterprise scheduling agent.

3. The job scheduling device according to Claim 2, wherein said job histories include information received from each enterprise scheduling agent regarding status of the jobs submitted.

4. The job scheduling device according to Claim 2, wherein said job data management device is utilized by said job scheduler to set parameters in jobs to be submitted to said enterprise scheduling WO 00/38033 PCT/US99/31024 d; c, 0 81 agent. ci The job scheduling device according to Claim 1, further comprising: 5 a job history repository that saves both jobs and C job histories of jobs submitted to each enterbrise C] scheduling agent; o wherein each enterprise scheduling agent C] comprises, an agent communicator configured to send and receive messages between said job scheduler and the enterprise scheduling agent, a job manager configured to setup, launch, run, and manage jobs submitted to the enterprise scheduling agent, a data manager configured to update and delete data from said job history repository, and a low level API configured handle internal functions of said enterprise scheduling agent (LES Agent), file management, and message handling functions.

6. The job scheduling device according to Claim 5, further comprising: an enterprise communicator configured to construct and communicate messages between said job scheduler and each enterprise scheduling agent; and a job data management device configured to maintain job histories of jobs submitted to each enterprise scheduling agent; wherein said data manager updates said job history via enterprise communicator messages sent from the enterprise scheduler to said job data management device. WO 00/38033 PCT1S9931024 82

7. The job scheduling device according to 0 Claim 1, further comprising: a command line device configured to accept commands regarding administration of jobs submitted to the enterprise scheduling agents; and Cl a job administration device confiau t Cl communicate said command line to at least one of said S enterprise scheduling agents for execution. Cl

8. The job scheduling device according to Claim 7, wherein: said commands accepted by said command line device include at least one of delete a job and all runs of the job, cancel a job's run, list all jobs, list all 1$ jobs by at least one of product code, status, and node, and rerun a job immediately.

9. The job scheduling device according to Claim 8, wherein: said commands accepted by said command line device include context variables; and said enterprise scheduling agent converts said context variables according to a current job and job parameters, and executes said commands. The job scheduling device according to Claim 1, further comprising: a point product device configured to provide a communication link between said enterprise scheduling agent and at least one product submitting jobs to said job scheduling device; wherein said point product device communicates job status, job logfile, setup, cancel, job parameter functions, and requests between each enterprise WO 00/38033 PCTMS99/31024 0 0 83 scheduling agent and said at least one product. C' II. The job scheduling device according t o Claim 10, further comprising: a job administration device configured to accept command line inputs and communicate said command line C] inputs to at least one enterprise scheduling agent; C a job data management device configured to o maintain job histories of jobs submitted to each enterprise scheduling agent; and an enterprise communicator configured to send messages between at least one of said job scheduler, point product device, job administration device, and job data management device and each of said enterprise scheduling agents.

12. The job scheduling device according to Claim 1, further comprising: an enterprise communicator configured to send messages between said job scheduler and each of said enterprise scheduling agents.

13. The job scheduling device according to Claim 12, wherein: each enterprise scheduling agent is registered at a specific node address that identifies each enterprise scheduling agent with a unique datagroup; and said enterprise communicator encodes each message with at least one destination corresponding to a datagroup to direct each message to at least one enterprise scheduling agent.

14. The job scheduling device according to Claim 1, further comprising: WO 00/38033 PCT/tS99/31024 ci .S 84 a local job repository installed on each of said C nodes; wherein: each local job repository maintains job and job S history information on each job submitted to the node where the local job repository is installed; ci each local job repository is updated by the O enterprise scheduling agent installed on the node where 0 the local job repository is installed; and said job information includes job parameters needed to execute each job. The job scheduling device according to Claim 14, further comprising: a job data management device configured to maintain job histories of jobs submitted to each enterprise scheduling agent; and a synchronizing device configured to synchronize each local job repository with the job histories maintained by said job data management device.

16. The job scheduling device accordin 3 to Claim 1, further comprising: a progress monitor configured to monitor and display execution of at least one of said jobs; wherein: said progress monitor provides a visual display of, an identification of said job and a current phase of said job, a percentage complete of said job, and a percentage complete of said current phase. WO 00/38033 PCT/US99'31024 0 O 0 85

17. The job scheduling device acccrdin to C-i Claim 1, further comprising: an autologin device configured to accept login parameters from a user submitting a job; 5 wherein said login parameters are utilized by an (C enterprise scheduling agent to launch and execute the job submitted. C]

18. The job scheduling device according to Claim 1, further comprising: a notification scripting device configured tc execute a notification script having instructions for notifying a user of status of a submitted job; wherein said notification scripting device includes facilities for creating, editing, and selecting a notification script for a specific job.

19. The job scheduling device according to Claim 1, wherein: said presentation layer includes, a GUI interface that accepts user inputs for scheduling and specifying a job to be submitted; wherein said GUI interface includes facilities for selection and creation of a scheduling calendar, selection of a start date and time, selection of recurring job run intervals, and selection of an immediate job run. The job scheduling device according to Claim 1, further comprising: a resource management device configured to enable a user to locate and view jobs and job runs. WO 00/38033 PCTIUS99/31024 ,l- 86 0

21. The job scheduling device according to 0- Claim 20, wherein: said resource management device includes an KM GI for defining an object representing a job, having, a general properties page having inpu fields for a label identifying the job, and a descripticn cf the job, o a description properties page having a selection 0 field for identifying an icon for representing the job, and a repository page having a selection field for identifying a time zone for display of job times.

22. The job scheduling device according to Claim 21, wherein: objects defined by said resource management device comprise, a hierarchy of folders including at least one of an all jobs folder, a jobs by group folder, a jobs by node folder, a jobs by product folder, a jobs by type folder, and a jobs by user folder.

23. The job scheduling device according to Claim 22, wherein said all jobs folder includes folders, including, an all jobs any status folder listing jobs regardless of status and associated job history of each job, an all runs by status folder listing jobs according to status, including completed runs, failed runs, not started runs, preempted runs, running runs, and stopped runs, a held jobs folder listing jobs that are held and can be scheduled for a later time, and W-O 00/38033 PCT/US99/31024 0 c, t 87- a scheduled jobs folder listing jobs that are 0- scheduled to run.

24. The job scheduling device according to C] 5 Claim 1, wherein: CJ said presentation layer includes, C]q a strategy scheduling window configured to allow o a user to view, create, modify, and delete schedules 0, for a strategy. A method of scheduling jobs across multiple networked computing platforms, comprising the steps of: determining at least one job based on job parameters for at least one job to be scheduled; sending said at least one job to at least one scheduling agent maintained on a selected nodes of said computer platforms; and executing each job on the selected node under management of said scheduling agent.

26. The method according to Claim 25, further comprising the steps of: monitoring progress of the executing job; and displaying said progress on a progress monitor.

27. The method according to Claim 25, further comprising the step of recording each job and a history of each job in a job history repository.

28. The method according to Claim 27, further comprising the step of: utilizing a job data management device for, WO 00/38033 PCTUS99/31024 0 88 retrieving status messages regarding each job sent from a scheduling agent of a selected node of said job, and updating said job history repository based on said C 5 status messages. c. ci

29. The method according to Claim. 28, further ocomprising the step of: 0maintaining a local job repositories, respectively on each of said nodes, each containing job and 4 ob history information for each job submitted to The respective node. The method according to Claim 29, further comprising the step of: synchronizing said job history repository with each local job repository.

31. The method according to Claim 25, wherein said step of determining comprises the steps of: retrieving said job parameters from one of a product and a user interface that collects said job parameters; validating said job parameters; and allocating a job based on said job parameters.

32. The method according to Claim 25, wherein said step of sending comprises the steps of: packaging said job parameters in a PEC communication format; and transmitting the packaged job parameters from a computing platform where said job parameters are determined to said scheduling agent maintain on the selected node. WO 00/38033 PCTfLS99!3I 024 0 ci 89 O

33. The method according to Claim 25, wherein said step of executing comprises the steps of: setting up the selected node to run an application program identified by said job parameters; i 5 executing said application program on the selected (c node; and CA monitoring progress of said application being o executed. ci

34. The method according to Claim 25, further comprising the steps of: accepting a command line for administration of jobs submitted to said enterprise scheduling agents; and communicating said command line to at least one of said enterprise scheduling agents for execution. The method according to Claim 34, further comprising the steps of: substituting context variables in said command line with data based on said context variable and the job to be administered; and executing the command line.

36. The method according to Claim 25, further comprising the step of: communicating data, including at least one of job status, job logfile, setup, cancel, job parameter functions, and requests for said data between a product and each enterprise scheduling agent.

37. The method according to Claim 25, further comprising the steps of: WO 00/38033 PCT/US99/31024 0 0 c, 0- 90 0 registering each enterprise scheduling agent at a C< node address that identifies the registered enterprise scheduling agent with a unique datagroup; communicating jobs and job administration commands and requests with each enterprise scheduling agent via C, messages; and encoding each message sent to a recipient O enterprise scheduling agent with at least one destination corresponding to a datagroup that directs said message to the recipient enterprise scheduling agent.

38. The method according to Claim 25, further comprising the steps of: retrieving autologin parameters from a user scheduling an autologin job; and launching execution of said job utilizing said autologin parameters.

39. The method according to Claim 38, further comprising the steps of: retrieving a notification script for a job being submitted; and executing the notification script on at least one of completion of said job and at a requested status point. The method according to Claim 25, further comprising the steps of: accepting a scheduling calendar identifying at least on of execution times and intervals for at least one of said jobs; and executing said jobs on selected nodes at the times and intervals identified in the calendar. W-O 00/38033 PCT/US99/31024 S- 91

41. The method according to Claim 25, further C comprising the steps of: providing a description of at least one cf said jobs, including a written description, a label, and an icon selected to represent said job; and C] identifying a time zone for display of job :imes. o 42. The method according to Claim 25, further 0- comprising the steps of: placing information about job times and status in an object containing folders, each folder identifying a categorization of jobs contained therein, including, an all jobs folder, a jobs by group folder, a jobs by node folder, a jobs by product folder, a jobs by type folder, and a jobs by user folder.

43. The method according to Claim 42, further comprising the step of: organizing said all jobs folder to maintain additional folders, including, at least one of, an all jobs any status folder listing jobs regardless of status and associated job history of each job, an all runs by status folder listing jobs according to status, including completed runs, failed runs, not started runs, preempted runs, running runs, and stopped runs, a held jobs folder listing jobs that are held and can be scheduled for a later time, and a scheduled jobs folder listing jobs that are scheduled to run.

44. The method according to Claim 25, further comprising the steps of providing a strategy scheduling WO 00/38033 PCT/US99/31024 O O 92 0 window that allows a user to view, create, modify, and 0 delete schedules for a strategy. A computer readable media, having 5 instructions stored thereon that, when loaded into a C computer, cause the computer to perform the steps of: C] determining at least one job based on job o parameters for at least one job to be scheduled; sending said at least one job to at least one scheduling agent maintained on a selected nodes of said computer platforms; and executing each job on the selected node under management of said scheduling agent.

46. A job scheduling device for scheduling jobs to run on at least one node of at least one computing platform, comprising: enterprise scheduling means installed on each node and configured to launch execution of jobs submitted to the enterprise scheduling means; presentation means configured to accept and validate parameters identifying at least one job to be submitted for execution on at least one of said nodes; and job scheduler means configured to allocate at least one job based on said parameters and submit the allocated jobs to at least one enterprise scheduling means. Dated this 19 th day of October 2004 COMPUTER ASSOCIATES THINK, INC SMOORENBURG PATENT TRADE MARK ATTORNEYS PO BOX 9 KANGAROO GROUND VIC 3097 AUSTRALIA