43

So I have a script to download a file from AWS daily and append it to a spread sheet. To do this I have set up a cronjob.

The Script works fine when I run it manually, but does not work when running from the cronjob.

The code has a line:

aws s3 cp s3://My/files/backup/ ~/home/AnPoc/ --recursive --exclude "*.tgz" --include "*results.tgz"

And in the email I recieve from the cronjob execution, I see the following error message:

./AnPoc/DayProcessing.sh: line 14: aws: command not found

I don't know why the command is not being found. Any help would be great.

AnPocArBuile
  • 635
  • 1
  • 7
  • 12

5 Answers5

59

First: check where on your system the executable aws is stored. Use this command:

$ which aws
/usr/bin/aws # example output, can differ in your system

Now, place a variable called $PATH in your crontab before the script:

PATH=/usr/bin:/usr/local/bin

Those paths separated by : define where should be search for the exectable. In the example above it's /usr/bin. You have to check all executables in your cron job that they are available.

Another thing: try to avoid path with a tilde (~) in cronjobs. Use /home/user instead.

chaos
  • 7,082
  • 3
  • 30
  • 38
  • 4
    If you `echo $PATH` you'll see that your session has the variable loaded. However when you execute from crontab it doesn't. – Ev. Sep 07 '16 at 09:01
  • 1
    Saved my ar$e... had to create and move backups of my mail server to s3 bucket. when I ran my script (with aws s3 cp ...) manually it worked, just adding the `PATH` was needed to make it work with cron. I swear if it were'nt for your post, I'd never be able to figure that out myself in another year. – Fr0zenFyr Nov 23 '16 at 17:42
  • 1
    @Ev. thanks for your comment, I noticed it even before the answer. It made me do a simple test by redirecting output of path to a file and the rest was explained in the answer above. – Fr0zenFyr Nov 23 '16 at 17:45
  • If you don't know, where to put in the crontab, see https://stackoverflow.com/a/2409369/1386952 – Christian Mar 18 '19 at 09:25
  • despite following all the requirements, still I am not able execute the cronjob when clubbing it with s3 command, can anyone please advice what am I doing wrong. `/usr/local/bin/aws s3 cp /home/ubuntu/sqlbackup.sql 's3://bucketName/db/sqlbackup.sql` – Sunil Kumar Feb 22 '22 at 11:39
34

You should use the full path for the aws command. For example, /usr/local/bin/aws

taras
  • 5,922
  • 10
  • 36
  • 48
libregeek
  • 1,244
  • 11
  • 23
  • 1
    @ user5735796 - when script is run through cron, it searches for aws executable in predefined locations as configured in $PATH variable. If aws is not installed in these locations, the command can't be found. So, either use the complete path to executable or add your aws location to $PATH variable. – plspl Mar 23 '17 at 03:15
10

Put this code before your command line to be executed into crontab -e

SHELL=/bin/bash
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
Roberto Carlos Reyes Fernandez
  • 571
  • 5
  • 4
3

The only thing that worked for me was to specify the path explicitly in the script:

ROOTDIR=/home/myusername
LOGDIR=$ROOTDIR/logs
DUMPDIR=$ROOTDIR/db_backup
LOGFILE=$LOGDIR/db_backup.log

$ROOTDIR/.local/bin/aws s3 cp $DUMPDIR/myappname-`date +"%Y-%m-%d"` s3://my-bucket-name/backups/myappname-`date +"%Y-%m-%d"` --recursive >> $LOGFILE 2>&1

As a previous poster said, use which aws to find the location of aws.

Little Brain
  • 2,208
  • 22
  • 45
0

If you have aws in your profile, you could also include your profile by adding . $HOME/.profile

* * * * * . $HOME/.profile; /path/to/command

Anthony Awuley
  • 2,835
  • 25
  • 18