top not working correctly

top not working correctly

Post by Oyvin Halfdan Thu » Sun, 27 Oct 2002 20:29:36



top reports everything just well, exept that both CPU load, usage,
system and user shows 0,0% all the time.

What's happening?

--
Oyvin

/*remove obvious parts of bogus reply-address
  to obtain working reply-address*/

 
 
 

top not working correctly

Post by Lowell Gilber » Mon, 28 Oct 2002 01:04:50



Quote:> top reports everything just well, exept that both CPU load, usage,
> system and user shows 0,0% all the time.

> What's happening?

Sounds like a bug.  
[There really isn't any more to say without full information.]

 
 
 

top not working correctly

Post by Lee Har » Mon, 28 Oct 2002 06:32:56



> top reports everything just well, exept that both CPU load, usage,
> system and user shows 0,0% all the time.

> What's happening?

Is it possible that your userland and kernel are not matched?
 
 
 

top not working correctly

Post by Simon Barne » Mon, 28 Oct 2002 06:43:52


Hi Oyvin,

Quote:> top reports everything just well, exept that both CPU load, usage,
> system and user shows 0,0% all the time.

Did you upgrade your source and build only the kernel, i.e. kernel and world are
out of sync now? It is important to avoid this, because top and serveral other
user land tools rely heavily on internal data structures and interfaces.

Regards,
 Simon

 
 
 

top not working correctly

Post by Oyvin Halfdan Thu » Mon, 28 Oct 2002 21:54:36



> > top reports everything just well, exept that both CPU load, usage,
> > system and user shows 0,0% all the time.

> Did you upgrade your source and build only the kernel, i.e. kernel and world are
> out of sync now? It is important to avoid this, because top and serveral other

No. It's 4.6.2 installed from CD not very long ago. It's a custom
kernel of course, but it have worked earlier. Actually I don't know if
I did anything special before i stopped working.

What proc files does top read for instance?

--
Oyvin

/*remove obvious parts of bogus reply-address
  to obtain working reply-address*/

 
 
 

top not working correctly

Post by David Malo » Mon, 28 Oct 2002 22:07:58



Quote:>What proc files does top read for instance?

It gets most of its information from sysctl and libkvm.

I'd suggest checking out "vmstat -i" and see if the rtc and clk
interrupts are working normally - the rates should be about 127 and
100 respectively.

There seems to be some problem with rtc interrupts getting lost on
some systems. Some of it may be due to suspend/resume problems on
laptops, so commenting apm out of the kernel config file (rather
than just disabeling it) may help.

        David.

 
 
 

top not working correctly

Post by jp » Mon, 28 Oct 2002 22:17:41


On 27 Oct 2002 13:54:36 +0100,

[snip]

Quote:

> What proc files does top read for instance?

None. It reads /dev/kmem and /dev/mem.

--
  j p d (at) d s b (dot) t u d e l f t (dot) n l .

 
 
 

top not working correctly

Post by Oyvin Halfdan Thu » Tue, 29 Oct 2002 00:58:44



> I'd suggest checking out "vmstat -i" and see if the rtc and clk
> interrupts are working normally - the rates should be about 127 and
> 100 respectively.

It is not reporting rtc, only clk. clk rate is 99.

Quote:> There seems to be some problem with rtc interrupts getting lost on
> some systems. Some of it may be due to suspend/resume problems on
> laptops, so commenting apm out of the kernel config file (rather
> than just disabeling it) may help.

This is a server machine, and as such, it's got a rather stripped-down
kernel with only the really necercary things compiled-in (alas: no apm).

--
Oyvin

/*remove obvious parts of bogus reply-address
  to obtain working reply-address*/

 
 
 

top not working correctly

Post by David Malo » Tue, 29 Oct 2002 05:41:27




>> I'd suggest checking out "vmstat -i" and see if the rtc and clk
>> interrupts are working normally - the rates should be about 127 and
>> 100 respectively.
>It is not reporting rtc, only clk. clk rate is 99.

Weird - that's the problem then - no rtc interrupts. I'd guess the
fact that it shows nothing means that they haven't been working
since boot time. Does this always happen, or just after some boots?

        David.

 
 
 

top not working correctly

Post by F. Liekwe » Tue, 29 Oct 2002 18:58:23


[top shows 0% cpu]

Hi, Oyvin,

  this symptom seems to occur when APM is in use.
I've seen reports that it occurs as soon as APM is
in the kernel;  I've definitely seen it happen on
my laptop running 4.6-RELEASE after putting it into
APM sleep mode.  My home machine (also 4.6-rel) has
APM in the kernel, and works fine, though.

Maybe try and take APM and related stuff out of Your
kernel, and see if it keeps happening.

If it does:  Maybe do a send-pr?

Yours,
Florian
--
=======================================================================
Florian Liekweg           | While any text editor can save your files,
IPD Universit?t Karlsruhe | only Emacs can save your soul.
==================================================[Per Abrahamsen]=====

 
 
 

top not working correctly

Post by Oyvin Halfdan Thu » Tue, 29 Oct 2002 22:15:37



> Weird - that's the problem then - no rtc interrupts. I'd guess the
> fact that it shows nothing means that they haven't been working
> since boot time. Does this always happen, or just after some boots?

I think it might be just after some boots. I hva seen it working earlier, but not recently (rebooted about 10 days ago). Do you think a reboot will do?

If so, are there any other ways to make it work on _every_ boot?

--
Oyvin

/*remove obvious parts of bogus reply-address
  to obtain working reply-address*/

 
 
 

top not working correctly

Post by David Malo » Tue, 29 Oct 2002 23:03:53




>> Weird - that's the problem then - no rtc interrupts. I'd guess the
>> fact that it shows nothing means that they haven't been working
>> since boot time. Does this always happen, or just after some boots?
>I think it might be just after some boots. I hva seen it working earlier, but not recently (rebooted about 10 days ago). Do you think a reboot will do?

If you can afford to reboot it a few times and see how often
it tends to happen, that might be interesting.

Quote:>If so, are there any other ways to make it work on _every_ boot?

Unfortunately, I don't think the problem is fully understood.
Most likely, the problem is related to twitchy hardware, but
it is difficult to say.

        David.

 
 
 

top not working correctly

Post by Oyvin Halfdan Thu » Wed, 30 Oct 2002 02:35:02



> >> Weird - that's the problem then - no rtc interrupts. I'd guess the
> >> fact that it shows nothing means that they haven't been working
> >> since boot time. Does this always happen, or just after some boots?

> If you can afford to reboot it a few times and see how often
> it tends to happen, that might be interesting.

I've rebooted the machine 10 times now without any luck :(

Quote:> >If so, are there any other ways to make it work on _every_ boot?

> Unfortunately, I don't think the problem is fully understood.
> Most likely, the problem is related to twitchy hardware, but
> it is difficult to say.

I've just put the dmesg out on http://www.oyvins.net/dmesg.txt if that's of
any interest.

--
Oyvin

/*remove obvious parts of bogus reply-address
  to obtain working reply-address*/

 
 
 

top not working correctly

Post by Lee Har » Wed, 30 Oct 2002 08:45:43


Quote:>   this symptom seems to occur when APM is in use.
> I've seen reports that it occurs as soon as APM is
> in the kernel;  I've definitely seen it happen on
> my laptop running 4.6-RELEASE after putting it into
> APM sleep mode.  My home machine (also 4.6-rel) has
> APM in the kernel, and works fine, though.

> Maybe try and take APM and related stuff out of Your
> kernel, and see if it keeps happening.

> If it does:  Maybe do a send-pr?

Except he said he is running a kernel without APM... hmm
Is there a BIOS setting do disable that stuff?
 
 
 

top not working correctly

Post by Jamie Jone » Mon, 04 Nov 2002 10:24:45


On 27 Oct 2002 20:41:27 -0000,

         wrote in newsgroup comp.unix.bsd.freebsd.misc:

Quote:>>> I'd suggest checking out "vmstat -i" and see if the rtc and clk
>>> interrupts are working normally - the rates should be about 127 and
>>> 100 respectively.

>>It is not reporting rtc, only clk. clk rate is 99.

> Weird - that's the problem then - no rtc interrupts. I'd guess the
> fact that it shows nothing means that they haven't been working
> since boot time. Does this always happen, or just after some boots?

Errrm, my PC doesn't show any rtc's either, and that's working fine!
Simliarly, my "clk" is 99 or 100 too.

(Athlon XP 2100+ / Abit KR7A-133 / 512 Mb / FreeBSD 4.7-RELEASE)

--
Jamie Jones, Gower, South Wales, UK.   http://www.bishopston.com/jamie/
---- 53 days to Christmas!        Word of the day: "trichocephaliasis"
----- "I'm not big, and I'm not clever - and I'm definitely not funny."
------- The valid reply address on this posting expires in 7 days time.

 
 
 

1. Getting top 3.5beta9 to work correctly with 64 bit executables

All,

Thanks for any help.  I discovered that my version of top (3.5beta9) can
no longer monitor the memory consumption of my 64 bit Oracle
executables.  It appears that my version of top has been compiled as a
32 bit app, and I suspect that it's having trouble reading the /proc
stuff correctly.

Does anyone know:

1) if recompiling top as a 64 bit app will help
and
2) if so, how to do it?  I tried to figure it out from the cc and ld man
pages and could get no further than "wrong machine class" on top.o on
the link step.  I figure that I'm close, but....

--
Wesley Gamble
Bison Consulting
(on site at The Men's Wearhouse)

Sent via Deja.com http://www.deja.com/
Before you buy.

2. GCC.bff for 4.2?

3. xlock not working correctly "Can not get hostbyname"

4. Q: usb-storage autoload (Toshiba PDR-M5)

5. scsi cdrom not working correctly

6. PCI messages

7. NFS not working correctly

8. IP accounting and NATD

9. su not working correctly (screwed up bigtime) - help !?

10. "su" not working correctly

11. ConfigXF86 does not work correctly, can't run X.

12. Adaptec 2940AU does not work correctly with OS5

13. udp time service not working correctly