Parallels Plesk 11 problems with error log rotation

[desingenia]

Hi, we have discovered problems with error log file rotation process on our instance of Parallels Plesk 11 that we have started running just a short time ago.

We actually discovered the problem when our disk space exceeded it's limits, from 15% to 100% of 25GB in less than 24 hours. After reviewing file directory we see that we have two error log files in /log diretory, one error_log.1 with over 6GB and another error_log with over 12GB and increasing at at a rate than more than 1GB per hour.

The content of such large files was due to warning logs originated from the website from internal processes (which are temporarily paused), nevertheless this is not the core problem, but error log files that are not rotating as they should be expected to. We actually expect Parallels Plesk to manage whichever amount of error logs, with complete independence of the ratio of which these error messages can or could be generated.
In Domain configuration -> Logs -> Log rotation it is configured as by default, with log maximum size as 10240 KB, and 10 maximum per day, but for some reason this is not working as it should.

We have also reviewed other posts with what seem to be similar issues to this, although we have not encountered any straight forward solution that solves something that in first instance seems to be a basic functionality.

We would thank any help from anybody that has had to mange with this issue, since we suppose we are not the first that have encountered this. Right now it is impossible for our customer to have a productive environment with this website, so any help would be really appreciated!!

Thanks in advance!! Best regards!!
Adrian.

 

[abdi]

Hope this discussion helps:

http://www.linkedin.com/groups/Problem-Plesk-Panel-1044-log-2892007.S.208119020

 

[desingenia]

OK, thanks for the hint. In our case we have to understand it was due to error_log file being over 12GB and Plesk not being able to compress this file.
After looking at both error_log and error_log.1 files, we see that error_log.1 is an exact copy of error_log only truncated at a certain register, so we understand this may be the file that is originated by Plesk intending to compress error_log file, mostly after experiencing that error_log.1 was created once again by Plesk just several hours after being deleted off the server, only that this time the file size was 8GB instead of 6GB.
In the end we renamed original error_log file so a new file has been created, and deleted error_log.1 from the server. For the last days no problem, although we'll wait for the error_log file to increase and watch on zipping processes.

Is this issue with large size log files a known bug from Plesk? Is there any official fix?

Thanks a lot and best regards, Adrian.

 

[abdi]

The problem usually occurs only on 32-bit systems, so in the long term, the best option would be to upgrade to a 64-bit environment.

More here:
http://kb.parallels.com/en/115572

 

[desingenia]

Thanks for the information Abdi !!