I have a php script I want to run every minute to see if there are draft news posts that need to be posted. I was using "wget" for the cron command in cPanel, but i noticed (after a couple days) that this was creating a blank file in the main directory every single time it ran. Is there something I need to stop that from happening?
Thanks.
When wget runs, by default, it generates an output file, from what I need to remember.
You probably need to use some option of wget, to specify to which file it should write its output -- and use /dev/null as destination file (It's a "special file" that will "eat" everything you can write to it)
Judging from man wget, the -O or --output-file option would be a good candidate :
-O file
--output-document=file
The documents will not be written to the appropriate files, but all will be concatenated together and written to file.
so, you might need to use something like this :
wget -O /dev/null http://www.example.com/your-script.php
And, btw, the output of scripts run from the crontab is often redirected to a logfile -- it can always help.
Something like this might help, about that :
wget -O /dev/null http://www.example.com/your-script.php >> /YOUR_PATH_logfile.log
And you might also want to redirect the error output to another file (can be useful, to help with debugging, the day something goes wrong) :
wget -O /dev/null http://www.example.com/your-script.php >>/YOUR_PATH/log-output.log 2>>/YOUR_PATH/log-errors.log
Related
First of all, I would like to point out that this is NOT a duplicate to this or this and other related questions. I am asking how to get the output as well and my question is mostly related to that.
I need to run a bash script from PHP and get its output in PHP. However, doing:
echo shell_exec('script');
dies out on the 60th second due to Apache's FcgidIOTimeout. If I use the following to run it in the background, I won't be able to get the output:
shell_exec('<myscript> > /dev/null 2>/dev/null &');
I cannot:
Modify the bash script;
The code and all the functionality needs to be in a single file (unless I use a temp file);
If I were to write the output of the shell_exec() function to a temp file, I would also need to find a way to verify whether the
process has finished;
Using a database is an overkill which I cannot afford;
My current and only idea is to create a tmp file using PHP and write the output to it, then using top -c | grep <myscript> (with a refresh button) and if it returns true then it is still ongoing. However, I suspect that would not be practical and efficient in most of the time.
It is important that I use a temp file and not actually creating a "permanent" file to write to.
Solution for a similar problem: A few years ago I had a similar problem. I had to upload a file to a FTP server. I wondered how to communicate to the FTP server that the file was uploaded completely, so that the FTP server can perform some tasks on it. My solution was to rename the file to something like *.completed after it was uploaded completely. Then the process on the FTP server could look for *.completed files only.
Solution adjusted to your problem: I'd suggest you to rename your temp file after it was generated by your bash script. This way you can independently find out if the script was executed successfully. shell_exec() could look like this:
shell_exec('<myscript> > tmp-file && mv tmp-file tmp-file.completed &');
Be aware that this only redirects the channel STDOUT into the the tmp-file. If you also want to redirect STDERR into the tmp-file, try this:
shell_exec('<myscript> &> tmp-file && mv tmp-file tmp-file.completed &');
In a apcahe server i want to run a PHP scripts as cron which starts a php file in background and exits just after starting of the file and doesn't wait for the script to complete as that script will take around 60 minutes to complete.how this can be done?
You should know that there is no threads in PHP.
But you can execute programs and detach them easily if you're running on Unix/linux system.
$command = "/usr/bin/php '/path/to/your/php/to/execute.php'";
exec("{$command} > /dev/null 2>&1 & echo -n \$!");
May do the job. Let's explain a bit :
exec($command);
Executes /usr/bin/php '/path/to/your/php/to/execute.php' : your script is launched but Apache will awaits the end of the execution before executing next code.
> /dev/null
will redirect standard output (ie. your echo, print etc) to a virtual file (all outputs written in it are lost).
2>&1
will redirect error output to standard output, writting in the same virtual and non-existing file. This avoids having logs into your apache2/error.log for example.
&
is the most important thing in your case : it will detach your execution of $command : so exec() will immediatly release your php code execution.
echo -n \$!
will give PID of your detached execution as response : it will be returned by exec() and makes you able to work with it (such as, put this pid into a database and kill it after some time to avoid zombies).
You need to use "&" symbol to run program as background proccess.
$ php -f file.php &
Thats will run this command in background.
You may wright sh script
#!/bin/bash
php -f file.php &
And run this script from crontab.
This may not be the best solution to your specific problem. But for the record, there is Threads in PHP.
https://github.com/krakjoe/pthreads
I'm assuming you know how to use threads, this is very young code that I wrote myself, but if you have experience with threads and mutex and the like you should be able to solve your problem using this extension.
This is clearly a shameless plug of my own project, and if the user doesn't have the access required to install extensions then it won't help him, but many people find stackoverflow and it will solve other problems no doubt ...
How to setup a cron job command to execute an URL?
/usr/bin/wget -q http://www.domain.com/cron_jobs/job1.php >/dev/null 2>&1
Why can't I make this work!? Have tried everything.. The PHP script should send an email and create some files, but none is done
The command returns this:
Output from command /usr/bin/wget -q http://www.domain.com/cron_jobs/job1.php ..
No output generated
... but it still creates an empty file in /root on each execute!? Why?
Use curl like this:
/usr/bin/curl http://domain.com/page.php
Don't worry about the output, it will be ignored
I had the same problem. The solution is understanding that wget is outputting two things: the results of the url request AND activity messages about what it's doing.
By default, if you do not specify an output file, it will create one, seemingly named after the file in your url, in the current folder where wget is run.
If you want to specify a different output file:
-O outputfile.txt
will output the url results to outputfile.txt, overrwriting what's there.
If you wish to append to that file, write to std out and then append to the file from there:
and here's the trick: to write to std out use:
-O-
the second dash is in lieu of a filename and tells wget to write the url results to std out.
then use the append syntax, >>, to append to a file of your choice:
wget -O- http://www.invisibility.com >>/var/log/invisibility.log
The lower case o, specifies the location of the activity log, so if you wish to log activity for the url request, you can:
wget -o http://someurl.com /var/log/activity.log
-q suppresses output of activity messages
wget -q http://someurl.com /var/log/activity.log
will not log any activity to the specified file, and I think that is the crux where people get confused.
Remember:
-O is shorthand for --output-document
-o is shorthand for --output-file, which is the activity log.
Took me hours to get it working. Thank you for people writing down solutions.
One also needs to make sure to check whether single or double quotes are needed, otherwise it will parse the url wrong leading to error messages:
This worked (using single quotes):
/usr/bin/wget -O -q 'http://domain.com/cron-file.php'
This gave errors (using double quotes):
/usr/bin/wget -O -q "http://domain.com/cron-file.php"
Don't know if the /usr/bin/ is needed. Read about different ways of how to do the order of the -O -q. It is hard to find a reliable definitive source on the web for this subject. So many different examples.
An online wget manual can be found here, for the available options (but check with the Linux distro one is using for an up to date version):
http://unixhelp.ed.ac.uk/CGI/man-cgi?wget
For use wget to display HTML:
wget -qO- http://www.example.com
I have a cron command below that is sending me status outputs on the wget.... I really don't want these outputs, I just want the code to run.
wget http://www.domain.com/cron/dailyEmail 2>&1;
How can I turn off the output?
Send the output to the null device.
wget http://www.domain.com/cron/dailyEmail >/dev/null 2>&1
Send it to a temporary file thus:
wget http://www.domain.com/cron/dailyEmail >/tmp/my_wget.out 2>&1
That way, you can see the output if you need to but it doesn't otherwise bother you.
If you want to keep older copies of the output around rather than over-writing them on each run, you can use something like:
wget http://www.domain.com/cron/dailyEmail >/tmp/my_wget_$(date +%Y_%m_%d).out 2>&1
which will give you a filename containing the date (and time if you change the arguments to the date command) but then you'll probably want an automated process to clean up older log files.
I'm trying to get a cron job to run every 5 min on my localhost. Using the Cronnix app I entered the following command
0,5 * * * * root curl http://localhost:8888/site/ > /dev/null
The script runs fine when I visit http://localhost:8888/site/ in my browser. I've read some stuff about getting CI to run on Cron, using wget and various other options but none make a lot of sense.
In another SO post I found the following command
wget -O - -q -t 1 http://www.example.com/cron/run
What is the "-O - -q -t 1" syntax exactly?
Are there other options?
-O - Means the output goes to stdout (-O /dev/null) would nullify any output. -q means be quiet (don't print out any progress bars), this would screw up the look of any log files. -t 1 means to only try once. If the connection fails or times out it will not try again.
See http://linux.die.net/man/1/wget for a full manual on the wget command.
Edit: just realised you're piping all this to /dev/null anyway, you may as well either omit the -O parameter or point that to /dev/null and omit the final pipe.
What I always do is use PHP in cli mode. Seems more efficient to me.
first setup a cron entry like :
*/5 * * * * /usr/bin/php /var/www/html/cronnedscript.php
cronnedscript.php should be placed in your root www folder.
then edit cronnedscript.php with:
<?php
$_GET["/mycontroller/index"] = null;
require "index.php";
?>
where mycrontroller is the CI controller you want to fire.
if you want the controller to only be run by crond ,as opposed through public www requests, add the following line to the controller and to the cronnedscript.php :
if (isset($_SERVER['REMOTE_ADDR'])) die('Permission denied');
I realize that this is a reference to Drupal, however they do a very nice job of explaining what each and every parameter is in the wget call.
Drupal Cron Explanation
If you want the more generic explanation, you can find it here.
Try this and save it by making a folder in the C drive with a .bat extension.
Then give the path of this script to task scheduler.
Then run the same.
C:\xampp\php\php-win.exe -fC:\xampp\htdocs\folder name\index.php controllername functionname