Does the quota limit also limit processes from other users like apache?

[Thomas Becker]

Hello,
does the quota limit also limit processes from other users like apache?

The reason why I'm asking this is because of the problem that the log files (i.e. error_log) sometimes are growing extremely fast and filling up the whole disk in a few hours. If I'm configuring the quota, is it still possible that the log files which are stored on each Webspace can grow higher than the quota for that Webspace account?

OS: CentOS 6

Regards,
Thomas

 

[InderS]

Hello,

Quota setup dose not setup the limit on any user process.

Can you please let me know which error logs files are using lot of disk space on your server ? It's server logs file OR any particular domain error_log file

Also I will suggest you check that file logs and try to resolve that issues so that you will not face any issues with the disk space

 

[Thomas Becker]

Hello InderS
Thanks for your answer.
It is the error_log of some domains in Plesk. It occurs sometimes and on different domains. I never checked an error_log which was that large because I do have to delete it as fast as possible to free up disk space. Any idea why the error_log is sometimes growing that fast without rotation and how to avoid such behavior?
Thanks.

Regards Thomas

 

[InderS]

Hello,

May be there is an issues with your PHP OR Apache configuration and due to that you are domain's error log files is growing very fast. You will have to check that error log files. You can check that using the tail commend so that you will get the live logs on your screen, Please check that and update here so that I can assist you in this regard


Thanks

 

[Thomas Becker]

Hello InderS,
thanks again for your answer.

There are no considerable errors in the log file at the moment (and also in the past). But sometimes the log is growing that fast within hours that the disk is completely full. There are definitely no errors which are present continuously. I have activated quota and will watch the behavior. I will update the threat if I get new information.

Regards
Thomas

 

[InderS]

You are most welcome

Please update here when you found the new information on this.

 

[Thomas Becker]

Hello InderS,
I recently got again a huge error_log (over 5 GB). Then the log was compressed (30 MB).
In the log there are millions of entries (php warnings). In this case it’s a Wordpress plugin causing a loop or something similar. The Google bot was viewing the content as I can see based on the IP address.

Error message:
PHP Warning: next() expects parameter 1 to be array, integer given in …

I'm sure the log rotation sometimes does not work and the log file is filling up the whole disk. In this case the log rotation started only when the error_log was greater than 5 GB. I set up log rotation in Plesk for files that are greater than 10 MB. It seems that the log rotation sometimes is not working.

There is a problem with error_logs sometimes filling up the whole disk for years now.
Are there any possibilities to prevent from such behavior?

CentOS 6.4
Kernel: 2.6.32
Plesk 10.4.4

Thanks for help.

 

[Thomas Becker]

I think the problem could be that the log rotation is only performed once a day, which is too seldom for fast growing log files.

As the Plesk script is executed daily and is not only rotating log files, it is necessary to have a script and a cronjob that can be setup to run hourly.

Is it right that the following script that is executed daily by root is responsible for log rotation?:
/usr/local/psa/libexec/modules/watchdog/cp/pack-sysstats day

Is there a script or any other solution for an efficient script that is only rotating the log files?
I’m pretty sure this will fix the issue for me and many others that have this problem.

 

[Thomas Becker]

Hello,
I worked out a solution.
The logrotation of Plesk is performed daily or weekly which is too seldom.

To run the logrotation manually:
/usr/local/psa/logrotate/sbin/logrotate /usr/local/psa/etc/logrotate.conf

So I wrote a little bash script that can be executed for example hourly:

#!/bin/bash
if /bin/find /var/www/vhosts/*/statistics/logs/ -type f -size +500M -exec /usr/bin/du -h {} \; | /bin/grep error_log
then
echo -e "\nStarting the compression of the log files (logrotate)..."
/usr/local/psa/logrotate/sbin/logrotate /usr/local/psa/etc/logrotate.conf
fi

The script is listing log files that are bigger than 500 MB with its size if there are some and is then executing the logrotation.