Description
Objectives
To implement a simple job scheduler that executes non-interactive jobs (e.g., jobs that do not have direct user interaction, jobs that can run in the background).
Description
In this project we will implement a simple job scheduler that will execute non-interactive jobs (for example, jobs that do not have direct user interaction, jobs that can run in the background).
At any given time, only P jobs should be executing, and P is provided as an argument to your program.
If you have more than P jobs submitted, then these additional jobs must wait until one of the P executing jobs are completed.
You can assume that P is typically the number of cores that are available on a system and you are executing one process per core and if there are more jobs than the available number of cores these jobs must wait for one of the processes to complete.
When you launch the program with P as the command-line argument, the program should print a prompt and wait for the user to enter commands.
As the jobs are non-interactive, the output and error streams from the job must be written to separate files – <jobid>.out and <jobid>.err, where <jobid> is the appropriate job id that is associated with a job.
The following commands must be supported by your program:
Commands
submit program arguments
showjobs
Description
Create a new process to execute the program specified with any arguments and print a jobid to standard output.
List all process that are either currently waiting or running (only those process that were started using the submit command). The output should include the jobid assigned to each job and the status of the jobs (whether it is running or waiting). If the job has completed, it will not be listed.
List all the processes that were executed by your job scheduler, including the name
submithistory of the job, the jobid that was assigned to it, the start time and end time of the job, and the status of the job (whether the job completed successfully or not).
• submithistory command is required only for CS 532 students.
Example
The following is a sample session:
$ ./mysched 2
Enter command> submit /home/UAB/unan/CS332/shared/hw1 1000 job 1 added to the queue
Enter command> submit /home/UAB/unan/CS332/shared/hw2 1000 job 2 added to the queue
Enter command> submit /home/UAB/unan/CS332/shared/hw3 1000 job 3 added to the queue
Enter command> showjobs
jobid
command
status
1
/home/UAB/unan/CS332/shared/hw1 1000
Running
2
/home/UAB/unan/CS332/shared/hw2 1000
Running
3
/home/UAB/unan/CS332/shared/hw3 1000
Waiting
Enter command>
#after some time
Enter command> showjobs
jobid
command
status
1
/home/UAB/unan/CS332/shared/hw3 1000
Running
Enter command> submithistory
jobid
command
starttime
endtime
status
1 /home/UAB/unan/CS332/shared/hw1 1000 Thu Oct 10 17:43:44 2019 Thu Oct 1
0 17:43:44 2019 Success
3 /home/UAB/unan/CS332/shared/hw2 1000 Thu Oct 10 17:43:44 2019 Thu Oct 1
0 17:43:44 2019 Success
Enter command>
Guidelines and Hints
1. Do not hardcore the value of P, make sure that your program will work for any value of P that can be between 1 to maximum number of cores available on a system (say, 8). You must test your program with at least P = 1, 2, 4 on the CS Linux servers.
2. You will need a queue to keep track of the jobs that are submitted. You can use the sample circular queue implementation provided here: queue.tar
3. You can work on this project with a partner (group of two students at most), or you can do it alone if you choose. You can
Submission
You are required to submit the lab solution to Canvas.
To upload solution to Canvas: Upload the C source code, PDF version of C source code, and README.md file to Canvas. Do not include any executable or object files.
Use the following command to create PDFs of your source code.(replace
the <Source_code_File_name> and <Output_Source_code_File_name> with your C source code file name):
$enscript <Source_code_File_Name>.c -o – | ps2pdf – <Output_Source_code_File_ Name>.pdf
****GitHub Submission is also required for this homework (If you are working with a partner, we want to see each of your contribution through github)
Grading Rubrics
Parsing and interpreting the commands (including command-line arguments)
Implementing the job scheduler (including the queue)
Implementing the commands
Documentation (README.md)
Total
20 points
40 points
30 points
10 points
100 points
