Excessive resource usage > uptime...

[markh]

I have a new VPS hosted on an openvz node. I'm getting loads of Excessive resource usage / CPU time emails. E.g.

Code: Select all

Time:         Mon Mar 10 15:44:59 2014 -0400
Account:      syslog
Resource:     Process Time
Exceeded:     273300 > 1800 (seconds)

Executable:   /sbin/syslogd
Command Line: /sbin/syslogd -u syslog
PID:          410 (Parent PID:410)
Killed:       No

I know how to cancel these messages, but the CPU time is a little over 189 days, on a server that has been running for 2 days.... is csf somehow grabbing data from the kernel that doesn't relate to my container... ?

Here a (partial) ps list showing syslogd with a tiny CPU usage.

Code: Select all

# ps ax
  PID TTY      STAT   TIME COMMAND
    1 ?        Ss     0:04 init
    2 ?        S      0:00 [kthreadd/8136]
    3 ?        S      0:00 [khelper/8136]
   69 ?        S      0:00 upstart-udev-bridge --daemon
   91 ?        Ss     0:07 /usr/sbin/sshd -D
   96 ?        Ss     0:00 /sbin/udevd --daemon
  123 ?        Ss     0:00 dbus-daemon --system --fork --activation=upstart
  227 ?        S      0:00 /sbin/udevd --daemon
  228 ?        S      0:00 /sbin/udevd --daemon
  254 ?        S      0:00 upstart-socket-bridge --daemon
  298 ?        Ss     0:03 cron
  410 ?        Ss     0:08 /sbin/syslogd -u syslog
...

I should add I'm getting messages for apache preforks and others with similar characteristics

I prefer to understand what's happening before blindly ignoring messages.

[ForumAdmin]

(((273300/60)/60)/24) = 3.16 days. The time is taken from the PID info in /proc/PID/stat and converts the time from jiffies to seconds based on the reported clock ticks.

[markh]

Thanks for that. I read the email as CPU time, not elapsed time...