can't run wkhtmltopdf from a cronjob - linux

Question

I got a command line.. When running it from putty it works, but when running the command from a cronjob (webmin) running as root the command hangs and never completes executing..
/usr/bin/xvfb-run -a -s "-screen 0 640x480x16" /usr/bin/wkhtmltopdf /root/input.html /root/output.pdf
update
Command line in cronjob.php
echo shell_exec('/usr/bin/xvfb-run -a -s "-screen 0 640x480x16" /usr/bin/wkhtmltopdf /root/input.html /root/output.pdf');
Command for the cron job (running as root)
php -f /var/cronjob.php
When the cron job is running from webmin the execution never completes, but when running the exact same command from putty it works! This is the output
Loading page (1/2)
Printing pages (2/2)
Done
Exit with code 1 due to network error: ProtocolUnknownError
Running the command (without wkhtmltopdf) from both putty and webmin works
echo shell_exec('/usr/bin/xvfb-run -a -s "-screen 0 640x480x16"');
this is the output
xvfb-run: usage error: need a command to run
Usage: xvfb-run [OPTION ...] COMMAND
Run COMMAND (usually an X client) in a virtual X server environment.
...
When adding wkhtmltopdf the cronjob never completes
update II
This command line doesn't work either from a cron job
xvfb-run -a -s "-screen 0 640x480x16" wkhtmltopdf -h
# Grokify
echo shell_exec('0 0 * * * * xvfb-run -a -s "-screen 0 640x480x16" wkhtmltopdf /var/www/tmp/test.html /var/www/tmp/output.pdf >> /var/www/tmp/pdf.log 2>> /var/www/tmp/pdf.err');
pdf.err
sh: 1: 0: not found

Answer 1

The cronjob may not be pulling in the environment of the user and therefor doesn't know what $PATH actually contains. I've found I need to use the full path to binaries in my crons:
2 * * * /usr/bin/php -f /var/cronjob.php

Answer 2

whats the user the cron is running as? it may be a permission issue. try to use sudo to give power to create the full file.
So in the cron have
sudo xvfb-run -a -s "-screen 0 640x480x16" wkhtmltopdf input.html output.pdf
you should consider putting xvfb-run in a full path like /usr/bin/xvfb-run or /bin/xvfb-run and when you specify input.html and output.pdf try to put its full path like /home/user/Documents/input.html /home/user/Documents/output.pdf

Answer 3

I'd assume problems with the (reduced) environment of cron and, subsequently, with xauth.
The way, I would try to gain progress is
(a) use the option --error-file=/tmp/xvfb.log of xvfb-run to see, what it says, and
(b) use the option --auth-file=/path/to/root_s_home/.Xauthority of xvfb-run
Another way round may be the installation of wkthtmltopdf from sources to get a real "headless" program without the need of xvfb-run
(These are just suggestions - I don't have chances to build a test scenario...)

Answer 4

The following PHP and crontab examples work for me both as unprivileged and privileged users. The example redirects STDOUT to /tmp/pdf.log and STDERR to /tmp/pdf.err to capture and examine logging and error messages. This example assumes xvfb-run and wkhtmltopdf are in your PATH which can be hard coded as well.
PHP
echo shell_exec('xvfb-run -a -s "-screen 0 640x480x16" wkhtmltopdf /var/www/tmp/test.html /var/www/tmp/output.pdf >> /var/www/tmp/pdf.log 2>> /var/www/tmp/pdf.err');
crontab
0 0 * * * xvfb-run -a -s "-screen 0 640x480x16" wkhtmltopdf /var/www/test.html /var/www/output.pdf >> /var/www/tmp/pdf.log 2>> /var/www/tmp/pdf.err
STDOUT: wkhtmltopdf Success
When run, the following appearing in /var/www/tmp/pdf.log indicating success which the error file remains empty:
$ cat /var/www/tmp/pdf.log
Loading page (1/2)
Printing pages (2/2)
Done
STDOUT: wkhtmltopdf Error
If there's a wkhtmltopdf error, it appears in STDOUT as well. The following is an example file not found error for test.html if the input file (/var/www/tmp/test.html) doesn't exist:
$ cat /var/www/tmp/pdf.log
Loading page (1/2)
Error: Failed loading page http:///var/www/tmp/test.html (sometimes it will work just to ignore this error with --ignore-load-errors)
[============================================================] 100%
Try adding the STDOUT and STDERR redirects to capture and check for error messages.