Home Linux Admin Introduction Directory Tree Disks/File Systems Memory Mgmt Monitoring Startup/ShutDown Logging in/out User Accounts Backups Processes Cron Packages Books

Scheduling processes Part 2

Contents

This chapter discusses scheduling in Linux.

anacron

We have seen how handy "cron" is at scheduling jobs. However there is a slight shortcoming to "cron" . Suppose we need to run something on a weekly basis and we schedule it for Monday morning 9am. At around 8am on Monday morning a storm comes and brings down the power. The power is restored at 10 am and cron does not run the job because the time has passed. This is where "anacron" comes in. It will run the job if the job was missed in the past. It is not a daemon but a program and works together with cron. The entries are kept in the file "/etc/anacrontab". However we do not have access to modify this file. We can take a look at the contents on the hills server. 
# /etc/anacrontab: configuration file for anacron

# See anacron(8) and anacrontab(5) for details.

SHELL=/bin/sh
PATH=/sbin:/bin:/usr/sbin:/usr/bin
MAILTO=root
# the maximal random delay added to the base delay of the jobs
RANDOM_DELAY=45
# the jobs will be started during the following hours only
START_HOURS_RANGE=3-22

#period in days   delay in minutes   job-identifier   command
1	5	cron.daily		nice run-parts /etc/cron.daily
7	25	cron.weekly		nice run-parts /etc/cron.weekly
@monthly 45	cron.monthly		nice run-parts /etc/cron.monthly
The entry consists of 4 fields. The first is the period in days. The "anacron" program does not have finer granularity. The second is minutes to wait for before running the job. The third field is the job name and the fourth is the command itself.

If the system is up then cron runs the job and informs anacron with the command "anacron -u jobname" . If however the system was down then when it comes up "anacron" will check the folder "/var/spool/anacron" for the timestamps of the jobs. On our hills server the folder looks like:

If it finds that the job has not been run then it will wait for the delay specified in the third fields , run the job and update the timestamp of the job.

nice

We can change the priority of a process with the "nice" command. This gives a hint to the CPU as to how much time should be spent on this process. We can see the priority of our currnet processes as below: 
[amittal@hills crontab]$ ps -el | more
F S   UID     PID    PPID  C PRI  NI ADDR SZ WCHAN  TTY          TIME CMD
4 S     0       1       0  0  80   0 - 62552 -      ?        00:28:29 systemd
1 S     0       2       0  0  80   0 -     0 -      ?        00:00:08 kthreadd
1 I     0       3       2  0  60 -20 -     0 -      ?        00:00:00 rcu_gp
1 I     0       4       2  0  60 -20 -     0 -      ?        00:00:00 rcu_par_
gp
1 I     0       6       2  0  60 -20 -     0 -      ?        00:00:00 kworker/
0:0H-events_highpri
The value appears under the "NI" column. Unfortunately we don't have permission to set the nice values for our processes on the hills server. The higher positive values means a higher priority. The command "renice" can change the priority of an existing process. 
nice -10 gnome-terminal

Negative value
nice --10 gnome-terminal

change priority of an existing process
sudo renice -n 15 -p 77982

run-parts

We noticed the phrase "run-parts" in the file "/etc/anacrontab" in the section on anacrontab. The "run-parts" is a command that can be used to run scripts in a folder, one by one. Create a folder ( mine is runparts1 ) and place the following 2 scripts in it. 
[amittal@hills runparts1]$ cat 1.sh
echo "1.sh" >> "/users/amittal/cs260a/crontab/runparts1/1.txt"
[amittal@hills runparts1]$ cat 2.sh
echo "2.sh" >> "/users/amittal/cs260a/crontab/runparts1/1.txt"

Change the full path of "1.txt" to be in the path of your new folder.
Change folder to 1 level above runparts1 and run the command "run-parts" .

[amittal@hills crontab]$ run-parts runparts1/
[amittal@hills crontab]$ cat runparts1/1.txt
1.sh
2.sh
[amittal@hills crontab]$

We print the contents of "1.txt" to show that all the scripts in that folder
have been run.
The "/etc/" has 3 folders: "/etc/cron.daily", "/etc/cron.weekly" , "/etc/cron.monthly" . As their name implies they run the scripts daily, weekly and monthly in their respective folders. This is a common scheme used in Linux systems. Now if we have a script that needs to runj daily we don't create an entry in the "anacrontab" file but instead drop the script in the "cron.daily" folder.

at

The utility "cron" and "anacron" allow us to execute a job at regular intervals. What if we wanted to execute a job only once. The utility "at" can be used for such a purpose. Example: 
Make sure at daemon is running ( ayd ) .
[amittal@hills crontab]$ ps -aef | grep "atd"
root        1642       1  0  2022 ?        00:00:00 /usr/sbin/atd -f
amittal  2909494 2233539  0 05:49 pts/1    00:00:00 grep --color=auto atd

[amittal@hills crontab]$ at -f sample.sh 07:00
warning: commands will be executed using /bin/sh
job 10 at Wed Jan  4 07:00:00 2023
[amittal@hills crontab]$ ls -ltr
total 20
-rw-r--r-- 1 amittal csdept  52 Dec 31 17:37 notes.txt
-rwx------ 1 amittal csdept  44 Dec 31 19:53 cp_sample.sh
-rwxrwxrwx 1 amittal csdept 225 Jan  2 07:27 sample.sh
-rwxrwxrwx 1 amittal csdept  30 Jan  2 18:45 sleep.sh
drwx------ 2 amittal csdept  70 Jan  3 22:48 runparts1
-rw------- 1 amittal csdept 145 Jan  4 07:00 1.txt
[amittal@hills crontab]$ cat 1.txt
Wed Jan  4 07:00:00 PST 2023
/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/sbin:/users/amittal/.local/bin:/users/amittal/bin
/users/amittal
[amittal@hills crontab]$
The command "at -f sample.sh 07:00" tells "at" to run the script "sample.sh" at 7:00 am. The "at" command can also control the jobs. We can list the jobs using the "jobq" or "job -l" command. After listing the jobs ( that are in the future... will not list past jobs ) we can choose to delete a job also with the "job -r" command. 
[amittal@hills crontab]$ at -f sample.sh 19:40
warning: commands will be executed using /bin/sh
job 12 at Wed Jan  4 19:40:00 2023
[amittal@hills crontab]$ atq
11      Thu Jan  5 07:00:00 2023 a amittal
12      Wed Jan  4 19:40:00 2023 a amittal
[amittal@hills crontab]$ at -l
11      Thu Jan  5 07:00:00 2023 a amittal
12      Wed Jan  4 19:40:00 2023 a amittal

[amittal@hills crontab]$ at -r 12
[amittal@hills crontab]$ atq
11      Thu Jan  5 07:00:00 2023 a amittal
The "at" command can take the time in different formats: 
    YYMMDDhhmm[.ss]. Specify an abbreviated year, month, day, hour, minute, and optionally seconds.
    CCYYMMDDhhmm[.ss]. Specify a full year, month, day, hour, minute, and optionally seconds.
    now. Indicates the current day and time and immediate execution.
    midnight. Indicates 00:00 AM.
    noon. Indicating 12:00 PM.
    teatime. Interpreted as 4 PM.
    AM. Indicates time before 12:00 PM.
    PM. Indicates time after 12:00 PM.
    today. The current day.
    tomorrow. The day after the current day.
Some examples executed on the hills server: 
Time already past . Assumes next day.

[amittal@hills crontab]$ at -f sample.sh 07:00
warning: commands will be executed using /bin/sh
job 14 at Thu Jan  5 07:00:00 2023

[amittal@hills crontab]$ at -f sample.sh -t 2304011950
warning: commands will be executed using /bin/sh
job 16 at Sat Apr  1 19:50:00 2023

[amittal@hills crontab]$ at -f sample.sh now
warning: commands will be executed using /bin/sh
job 17 at Wed Jan  4 19:53:00 2023


TO DO
1) Schedule a script to run at 8:30 pm today. If it's past 8:30 pm then
the script should run next dat. Use both "PM" and "20" hour notation to
schedule 2 jobs.
2) Schedule a script to run at 1:00 am on April 15th 2024 .
3) List the jobs and then delete the jobs.

at access

Access is done the same way as with cron. There are 2 files "at.allow" and "at.deny". On the hills server the file "at.allow" does not exist and the "at.deny" is empty. That means all users can use "at" .