Bug 76797

Summary:	top idle time incorrect
Product:	[Retired] Red Hat Linux	Reporter:	Need Real Name <pthompson>
Component:	procps	Assignee:	Alexander Larsson <alexl>
Status:	CLOSED DUPLICATE	QA Contact:	Brian Brock <bbrock>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	7.2
Target Milestone:	---
Target Release:	---
Hardware:	i686
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2002-10-31 16:09:40 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Need Real Name 2002-10-26 22:02:10 UTC

From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows 98; T312461)

Description of problem:
Idle time incorrect in top (procps-2.0.7-11)


Version-Release number of selected component (if applicable):


How reproducible:
Sometimes

Steps to Reproduce:
1.top
2.wait for less than 10 seconds on machine with load average over 1
3.watch idle time alternate between 0% and 8xxxxx.x%
	

Actual Results:  128 processes: 122 sleeping, 6 running, 0 zombie, 0 stopped
CPU states:  0.7% user,  2.1% system, 97.2% nice, 843804.7% idle
Mem:   255744K av,  250980K used,    4764K free,       0K shrd,   55260K buff
Swap:  554184K av,   26080K used,  528104K free                  108344K cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND


Expected Results:  0% idle

Additional info:

Linux 2.4.18-17.7.xcustom (root@malacandra) 
procinfo -f seems to display correct information

Comment 1 Marco De la Cruz 2002-10-31 16:09:34 UTC

I am running Red Hat 7.3. This is my experience with this bug:

I've recently upgraded to kernel-2.4.18-17.7.x.athlon.rpm (from
kernel-2.4.18-10.athlon.rpm) and I've noticed a problem with the
CPU idle %. Here is an example of what happens when running
"top":



 10:37am  up 8 days, 22:02,  9 users,  load average: 1.00, 1.01, 1.00
89 processes: 81 sleeping, 7 running, 1 zombie, 0 stopped
CPU states:  0.1% user,  0.1% system, 99.8% nice, 857278.7% idle
Mem:   514156K av,  485724K used,   28432K free,  0K shrd,   57956K buff
Swap:  522072K av,   44048K used,  478024K free             288448K cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
18840 mersenne  39  19 14996  14M   696 R N  99.9  2.9  7046m mprime 
14091 marco     15   0  1084 1084   848 R     0.1  0.2   0:00 top
    1 root      15   0   472  444   420 S     0.0  0.0   0:04 init
    2 root      15   0     0    0     0 SW    0.0  0.0   0:01 keventd
    3 root      15   0     0    0     0 SW    0.0  0.0   0:00 kapmd  
    
    
    
As you can see the idle value is absurd (it blows up about 1 in five
times, other updates are reasonable e.g. 0.2% idle). When running "top"
the value consistently jumps to about 800000% about 20% of the time.
If I run "top -d1" so that the updates take place every second the  
value consistently jumps to about 4200000%, again about 20% of the time.
Running "top -d2" makes it jump to around 2000000%. The idle value 
behaves approximately as follows:

        4200000
idle = --------% once every five updates.
       interval
       
If I stop "mprime" so that the system load becomes negligible the
idle value does not spike. If I run "mprime" or any other CPU-consuming
task (e.g. yes > /dev/null) the spikes commence. The niceness of
the process does not seem to matter.

My hardware is an Athlon 2100+ running on an ASUS A7V333.

> uname -a
Linux reimeika.math.toronto.edu 2.4.18-17.7.x #1
Tue Oct 8 11:49:30 EDT 2002 i686 unknown

> top --version
top (procps version 2.0.7)

> rpm -q procps
procps-2.0.7-12

Comment 2 Alexander Larsson 2002-11-05 10:01:17 UTC


*** This bug has been marked as a duplicate of 71237 ***