'sort' running server out of memory

[HostaHost]

Anyone from Parallels want to chime in on why servers are being run out of memory by the 'sort' command even when the logs are small? Here's what is currently on a site:

-rw-r--r-- 2 root root 22392143 Mar 26 11:40 access_log
-rw-r--r-- 2 root root 68018869 Mar 26 10:50 access_log.processed.1
-rw-r--r-- 3 root root 67671664 Mar 26 09:50 access_log.processed.2
-rw-r--r-- 1 root root 69908 Mar 26 10:50 access_log.stat
-rw-r--r-- 2 root root 45209044 Mar 26 10:50 access_ssl_log.processed
-rw-r--r-- 1 root root 0 Mar 26 10:50 access_ssl_log.stat
-rw-r--r-- 2 root root 355728 Mar 26 11:38 error_log
-rw-r--r-- 2 nginx root 19863420 Mar 26 11:40 proxy_access_log
-rw-r--r-- 2 nginx root 2238322 Mar 26 11:40 proxy_access_ssl_log
-rw-r--r-- 2 root root 0 Mar 26 10:50 xferlog_regular
-rw-r--r-- 2 root root 2357 Mar 26 10:50 xferlog_regular.processed
-rw-r--r-- 1 root root 0 Mar 26 10:50 xferlog_regular.stat

I have the site set to rotate at 100 MB, keep 1 file, and run

50 * * * * /usr/local/psa/admin/bin/statistics --calculate-one --domain-name=<domain>

hourly. Yet here's what is still happening:

Out of memory: Kill process 8569 (sort) score 879 or sacrifice child
Killed process 8569, UID 0, (sort) total-vm:5567932kB, anon-rss:3573156kB, file-rss:232kB
master invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0
master cpuset=/ mems_allowed=0
Pid: 1814, comm: master Not tainted 2.6.32-504.12.2.el6.x86_64 #1
Call Trace:

What is your software doing that sort needs 5 gigs of memory to complete?

See this unanswered thread:

http://talk.plesk.com/threads/webst...to-sort-command-on-access_log-webstat.325220/

 

[SibProgrammer]

Can you share some additional information?

Output of the commands (via ssh under root user):
1. plesk version
2. free -m

IP address of your server (to check error reports from your server on our side).

Typical size of files in logs directory (output of "ls -la" command).