Invalidly high process stats for some processes on 10.04
On 2 new 10.04 servers (at AWS) that we have recently set up, we get ridiculously high stats for a few processes, even immediately after reboot (this machine has only been up 22 hours at the moment). Looks like it only happens with mysqld and console-kit-daemon.
For example, from ps aux:
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
mysql 767 183065341 50.1 36772040 35958884 ? Sl Aug10 48408964:27 /usr/local/
root 936 208382845 0.0 188336 3996 ? Sl Aug10 46865656:43 /usr/sbin/
atop shows similar strangeness:
ATOP - perf1 2010/08/11 14:04:35 10 seconds elapsed
PRC | sys 5124095h35m | user 2351959867h05m | #proc 191 | #zombie 0 | #exit 0 |
CPU | sys 28% | user 62% | irq 0% | idle 673% | wait 36% |
CPU | steal 1% | stl/cpu 0% | | | |
cpu | sys 10% | user 18% | irq 0% | idle 55% | cpu000 w 16% |
cpu | sys 8% | user 15% | irq 0% | idle 75% | cpu001 w 1% |
cpu | sys 6% | user 11% | irq 0% | idle 79% | cpu002 w 4% |
cpu | sys 2% | user 7% | irq 0% | idle 90% | cpu003 w 1% |
cpu | sys 2% | user 6% | irq 0% | idle 92% | cpu004 w 0% |
cpu | sys 0% | user 2% | irq 0% | idle 86% | cpu005 w 12% |
cpu | sys 0% | user 2% | irq 0% | idle 98% | cpu006 w 1% |
cpu | sys 0% | user 1% | irq 0% | idle 99% | cpu007 w 0% |
CPL | avg1 4.63 | avg5 4.58 | avg15 3.76 | csw 119617 | intr 124132 |
MEM | tot 68.4G | free 6.8G | cache 21.5G | buff 216.6M | slab 163.2M |
SWP | tot 0.0M | free 0.0M | | vmcom 37.9G | vmlim 34.2G |
DSK | sdb | busy 40% | read 0 | write 937 | avio 4 ms |
NET | transport | tcpi 5226 | tcpo 4696 | udpi 0 | udpo 0 |
NET | network | ipi 5226 | ipo 4696 | ipfrw 0 | deliv 5226 |
NET | eth0 ---- | pcki 5167 | pcko 4637 | si 858 Kbps | so 6140 Kbps |
NET | lo ---- | pcki 60 | pcko 60 | si 2 Kbps | so 2 Kbps |
PID SYSCPU USRCPU VGROW RGROW RDDSK WRDSK ST EXC S CPU CMD 1/1
767 5124095h35m 2351959867h05m -34.1M 1508K 0K 872K -- - S 800% mysqld
31876 0.01s 0.41s 0K 0K 0K 0K -- - S 4% ruby
28193 0.01s 0.01s 0K 0K 0K 0K -- - R 0% atop
449 0.01s 0.00s 0K 0K 0K 8K -- - S 0% kjournald
5801 0.00s 0.00s 0K 0K 0K 0K -- - S 0% haproxy
159 0.00s 0.00s 0K 0K 0K 4K -- - S 0% kjournald
Any thoughts? We don't see this on a very similar 9.10 servers.
We don't do anything unusual in setup for these (e.g. they are stock 64-bit Amazon AMIs, not patched, etc). The only thing slightly unusual is that we use a third-party MySQL storage engine (tokutek), but they have never seen this issue either (although they said they don't do as much testing on Ubuntu--mostly CentOS and Fedora).
Thanks,
dwh
Question information
- Language:
- English Edit question
- Status:
- Answered
- For:
- Ubuntu Edit question
- Assignee:
- No assignee Edit question
- Last query:
- Last reply:
Can you help with this problem?
Provide an answer of your own, or ask Denis Haskin for more information if necessary.