Is it possible to make crontab run after one's execution. I read that cron jobs start from top, but run parallel, doesn't wait one to finish.
Is there any way that I can do like this?
For example, let's say I have 3 cron jobs.
*/5 * * * * job1
*/5 * * * * job2
*/5 * * * * job3
I want job2 to run after job1 has completed, and job3 after job2 completed.
You can have more than one command in a single crontab entry:
*/5 * * * * job1 ; job2 ; job3
In case you want to use the global crontab in /etc, it's simpler.
The main crontab file is /etc/crontab. It contains this type of instruction:
25 6 * * * root test -x /usr/sbin/anacron || ( cd / && run-parts --report /etc/cron.daily )
so Linux uses run-parts; run-parts launch commands in alphabetical order, so just add numbers in the script like this:
01_MyFisrtScrip
02_MySecondScript
...
and so on... Please avoid using dots or other characters that are not legal for run-parts:
BAD NAME = 01_myscript.sh
GOOD NAME = 01_myscript
Related
I made a script in TypeScript that download data from some api and store inside a mongo DB.
If i run yarn start from the app folder it works well.
I would like to put this command in a cron job that will be executed every 5 minutes.
I try it with some sintax in crontab but ti doesn't work.
I try to put the call in a run.sh script but it doesn't work too.
*/5 * * * * cd /opt/app-folder/src/ && /home/username/.nvm/versions/node/v16.15.1/bin/ts-node main.ts
*/5 * * * * cd /opt/app-folder && /usr/bin/yarn start > /home/username/app-name-out.txt
*/5 * * * * /home/username/run.sh > /home/username/app-name-out.txt
*/5 * * * * /home/username/.nvm/versions/node/v16.15.1/bin/ts-node /opt/app-folder/src/main.ts > /home/username/app-name-out.txt
*/5 * * * * cd /opt/app-folder/src/ && /home/username/.nvm/versions/node/v16.15.1/bin/ts-node main.ts > /home/username/app-name-out.txt
Can someone help me to execute the main.ts every 5 minutes?
Thanks
I get rid of this problem.
There was 2 problems, the first related to the output redirection.
I fixed by redirect stdout in a file and stderr in another one.
The second was related the the $PATH of crontab: it was /usr/bin:/bin.
To fix it I log into my user where script works and I print my $PATH with echo $PATH.
I copied the value and I set it before the crontab line in crontab file.
This is what it looks like:
# Set the same path of user username to have the correct path in script
PATH=/home/username/.nvm/versions/node/v16.15.1/bin:/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/snap/bin
# Execute oracle every 5 minutes
*/5 * * * * /bin/sh /home/username/run.sh >> /home/username/app-name-info.txt 2>> /home/username/app-name-error.txt
Now it works.
Some of my sites need regular crontabs, I use this to start a cronjob every 5 minutes "*/5 * * * *".
The crontabs are small, light, but there are starting to be several sites that need them, and starting them all together, it starts not being a very good idea.
With this "*/5" the cron starts at 5, 10, 15 20, etc... is it possible to make it start at, for example 8,13,18,23, etc?
Vixie cron accepts steps in a range (thanks Keith Thompson), so you can do
3-58/5 * * * * my_command
With other versions of cron, this may not be supported and you'd just have to do
3,8,13,18,23,28,33,38,43,48,53,58 * * * * my_command
Another option is something like
*/5 * * * * sleep 3m ; my_command
This could be adapted to sleep for a random time, thus further spreading out the jobs. For instance, you could do
*/5 * * * * /bin/bash -c 'sleep $((RANDOM/(32767/180))) ; my_command'
or use SHELL = /bin/bash further up in your crontab to make the /bin/bash -c unnecessary, if you're okay with using bash to run all the other cron jobs following the SHELL = line. $RANDOM in bash expands to a random integer between 0 and 32767, so this gets you a delay of up to 180 seconds.
I have a series of cron jobs running at the command line calling the php interpreter all by the same user configured to run once a day
0 0 * * * /usr/bin/php -q /mydirectory/myphp.php
0 0 * * * /usr/bin/php -q /mydirectory/myphp2.php
0 0 * * * /usr/bin/php -q /mydirectory/myphp3.php
Do these all execute at once or do the execute in the order of entry in some cron table, complete and move on to the next cron job?
And to answer your question despite the off-topic:
They will get executed in parallel, not sequentially. If you need some order it would pay to add them all to one script, and execute them sequentially separated by &&, e.g.
#!/bin/bash
/usr/bin/php -q /mydirectory/myphp.php && /usr/bin/php -q /mydirectory/myphp2.php && /usr/bin/php -q /mydirectory/myphp3.php
I have added the following entry:
*/1 * * * * /home/coddict/myapp-dev/spoolemailsender
and the shell that I am trying to execute (the file spoolemailsender) has the following:
#!/bin/sh
php app/console swiftmailer:spool:send --env=dev
Why isn't this script running every 1 minute? Do I need another command to get this cron job running?
You forgot to put user to execute cron job:
*/1 * * * * root /home/coddict/myapp-dev/./spoolemailsender
or
*/1 * * * * root sh /home/coddict/myapp-dev/spoolemailsender
root for example.
Assuming spoolemailsender is executable script and you don't need to do ./spoolemailsender or sh spoolemailsender
Is there a way to enable and disable Crontab tasks using Bash/Shell?
So when the user starts Server 1, it will enable the Server 1 Crontab line and so on.
And when the user stops Server 1, the Server 1 Crontab line get disabled (#).
Is this possible and how?
Thanks in advance
*/1 * * * * Server 1 check
*/1 * * * * Server 2 check
*/1 * * * * Server 3 check
SERVERNUM=$1
To enable:
crontab -l | sed "/^#.*Server $SERVERNUM check/s/^#//" | crontab -
To disable:
crontab -l | sed "/^[^#].*Server $SERVERNUM check/s/^/#/" | crontab -
Transcript:
barmar#dev$ crontab -l
*/1 * * * * Server 1 check
*/1 * * * * Server 2 check
*/1 * * * * Server 3 check
barmar#dev$ crontab -l | sed '/^[^#].*Server 1 check/s/^/#/' | crontab -
barmar#dev$ crontab -l
#*/1 * * * * Server 1 check
*/1 * * * * Server 2 check
*/1 * * * * Server 3 check
barmar#dev$ crontab -l | sed '/^#.*Server 1 check/s/^#//' | crontab -
barmar#dev$ crontab -l
*/1 * * * * Server 1 check
*/1 * * * * Server 2 check
*/1 * * * * Server 3 check
I suggest you add your cron jobs to /etc/cron.d for every server one script. Then let the cron script scan for some marker file if the cron job should be executed.
As a quick and dirty fix, you can enable or disable the execute permission of the appropriate cron script.
E.g. if you like to prevent locate from automatically updating its database (which can be I/O consuming):
cd /etc/cron.daily
sudo chmod a-x locate
This may be against the cron framework, but it is quickly applied and it works in case of immediate needs.
this is a variant, I use a cronjob that loads it self every night. I just edit a file and it gets reloaded at 10pm everynight. You could make the reload happen more often. I keep a directory of files for each of nodes. The trick is make sure that nobody comments out the reload line.
0 22 * * * crontab /home/ME/cron_files/NODE