[Laurence]

Recently, iTunes was crashing whenever I started it, rebooting did not help. I thought I'd just reinstall it and see if that made a difference. While in the optimizing phase of installation it caused a kernel panic. After rebooting the entire system is very unresponsive. I managed to get the terminal running and started top. It seems obvious that the reason the system is so unresponsive is that pid 0 kernel_task is using between 80 and 120% (how is 120% possible??) of the processor constantly. I left it running overnight and it wouldn't wake from sleep. Tried sshing in from a different computer and could not login (maybe it would although the ssh client times out before it gets a responce with a 5 minute timeout) I have rebooted a few times and it is the same every time. pid 0 always takes all the processor time it can. I can run applications, it just seems like its running on a 25MHz 68040 processor.

I did reinstall iTunes without crashing again and it didn't help. I also killed off all 3rd party processes (iTunes_helper, Microsoft_office, etc and it doesn't help.

Any ideas on where to go from here. I don't really want to reinstall completely if I don't have to.

Anyone?

[Jelle Monkmater]

Well, it sounds to me like you do have to reinstall. PID 0 showing is not supposed to happen. That is, I've never seen it, and if I would, I'd reinstall in an instant.

I don't know if it's useful to do a 'soft' reinstall as I don't know whether the kernel gets replaced during the 10.1 update, so you'd have to reinitialize your drive.

It's a pain, but so is a 25MHz processor.

[eno]

Originally posted by Jelle Monkmater:
<STRONG>Well, it sounds to me like you do have to reinstall. PID 0 showing is not supposed to happen. That is, I've never seen it, and if I would, I'd reinstall in an instant.

I don't know if it's useful to do a 'soft' reinstall as I don't know whether the kernel gets replaced during the 10.1 update, so you'd have to reinitialize your drive.

It's a pain, but so is a 25MHz processor.</STRONG>

Hmmm.... Well look at the following line from my "top" listing:

0 kernel_tas 6.5% 61:06.79 28 0 1367 7.66M+ 0K 63.0M+ 583M+

Darn! It looks like I'd better reinstall "in an instant"... so to speak...

And seeing as I don't know what happens to the kernel I better format my drive first....

See you guys when I get back!

[Jelle Monkmater]

Originally posted by eno:
<STRONG>

Darn! It looks like I'd better reinstall "in an instant"... so to speak...

And seeing as I don't know what happens to the kernel I better format my drive first....

See you guys when I get back!</STRONG>

Well I'll be damned. It doesn't show on my G4, but it shows on my iBook. I've not yet seen a UNIX system that shows PID 0, and I'm not at all sure why OSX shows it.

Anyway, I'm glad to see someone knows better and could come up with a proper solution to the problem at hand.

[aenemated]

holy shit ... i've got the same thing. here's from my top ...

0 kernel_tas 1.7% 1:29.88 26 0 - - - 41.5M 558M
0 idle_threa 57.9% 6:47:11

wtf is idle_threa that's eating 57.9% of my CPU?

[rantweasel]

Originally posted by aenemated:
<STRONG>holy shit ... i've got the same thing. here's from my top ...

0 kernel_tas 1.7% 1:29.88 26 0 - - - 41.5M 558M
0 idle_threa 57.9% 6:47:11

wtf is idle_threa that's eating 57.9% of my CPU?</STRONG>

Just to me too everybody, from my OpenBSD machine:
root 0 0.0 1.4 0 440 ?? DLs 18Jun01 1:05.77 (swapper)

and my OSX machine:
it's not there. But then, my OSX machine has no idle CPU, since it's running setiathome, and my OBSD machine does have idle CPU... Just in case I'm not responding to sarcasm, the idle_thread (and it's related threads or processes on other architectures and OSes) are null instructions executed by the CPU when it has nothing better to do...

[kangoo_boo]

Originally posted by aenemated:
<STRONG>holy shit ... i've got the same thing. here's from my top ...

0 kernel_tas 1.7% 1:29.88 26 0 - - - 41.5M 558M
0 idle_threa 57.9% 6:47:11

wtf is idle_threa that's eating 57.9% of my CPU?</STRONG>

bwahwahhaw
;p
kill -9 idle thread !!
lol

[Spheric Harlot]

Originally posted by rantweasel:
<STRONG>Just in case I'm not responding to sarcasm, the idle_thread (and it's related threads or processes on other architectures and OSes) are null instructions executed by the CPU when it has nothing better to do...</STRONG>

Ah - but why? isn't that just a waste of power? what's the benefit of doing this?

[lgerbarg]

Originally posted by Spheric Harlot:
<STRONG>

Ah - but why? isn't that just a waste of power? what's the benefit of doing this?</STRONG>

No, the idle thread usually tells the CPU to nap or doze when it is running. That means the CPU is using less power. You can implement really simple powersavings by having a low priority thread that just naps or dozes the CPU for a brief time whenever it is scheduled. Whenever there is nothing going on it gets scheduled more, and the thus you save more power.

Louis

[Spheric Harlot]

Originally posted by lgerbarg:
<STRONG>

No, the idle thread usually tells the CPU to nap or doze when it is running. That means the CPU is using less power. You can implement really simple powersavings by having a low priority thread that just naps or dozes the CPU for a brief time whenever it is scheduled. Whenever there is nothing going on it gets scheduled more, and the thus you save more power.
</STRONG>

So that 57% usage thread is actually a power-saving feature?

That's pretty cool - learning new things every day...

Thanks!

-chris.

[Laurence]

Back to the original post, I have discovered a little more information and it looks like it may be a bug in OS X. I decided to reinstall completely and formatted a 2nd drive, installed 10.0, then the 10.1 update and everything was fine. I then used software update to get all the new updates, and as soon as I installed iTunes 2.03 via software update the situation returned. idle_thread is 0 and kernel_task is hovering around 90-95%. It definitely is related to iTunes 2.03. Since many others are using the same version of iTunes, I assume it must be something about my system. It is a G4/533 with 1.5GB RAM and 2 40GB HDs. the only non-standard hardware is a que 24/10/40 CDRW drive installed in the internal drive bay. I haven't had a change to test without the drive, although I will swap it out to the original CD drive from the factory to test soon. With iTunes 2.02 this drive caused no issues. It worked fine under OSX using Toast Titanium preview 2 before the update. The strange thing is that I had iTunes 2.03 installed before this all started and although it didn't see the burner it did function properly otherwise. the only difference is the first time I installed iTunes, I used the standalone installer, not the software update veresion. Does anyone know if there is a difference or if SU just uses the same version? If anyone has any ideas as to what to do next, let me know. Obviously I will reinstall again and just use iTunes 2.02 until I figure this out.

[megus]

I've seen this kernel_task taking about 80-85% of the processor on a friends machine (beige G3 233Mhz). It would appear when he booted from Mac OS 9 into OSX. A new reboot OSX-&gt;OSX solved the problem. We could repeat this, so it's probably a bug in OS X.
I don't think it has anything to do with iTunes 2.03. My friend doesn't have it installed. I on the other hand has, on my b/w G3 and my iBook dual usb without having any problems. Maybe it's scsi related. He has a scsi cd-burner that doesn't work under OS X.

---
megus

[kangoo_boo]

Originally posted by lgerbarg:
<STRONG>

No, the idle thread usually tells the CPU to nap or doze when it is running. That means the CPU is using less power. You can implement really simple powersavings by having a low priority thread that just naps or dozes the CPU for a brief time whenever it is scheduled. Whenever there is nothing going on it gets scheduled more, and the thus you save more power.

Louis</STRONG>

that's probably right, but you are not explaining 'why``.
so why ?
because a cpu cannot just stop processing instructions. You must feed it, so you usually pass NOP (NoOPeration) which takes no real cpu power. you can also send powersaving stuff (idle bit), etc, as louis said.

[lgerbarg]

Originally posted by kangoo_boo:
<STRONG>

that's probably right, but you are not explaining 'why``.
so why ?
because a cpu cannot just stop processing instructions. You must feed it, so you usually pass NOP (NoOPeration) which takes no real cpu power. you can also send powersaving stuff (idle bit), etc, as louis said.</STRONG>

No, that is not what I meant, and that is not true. Many CPUs can stop processing instructions. From the PPC 7451 User Guide:

Nap�<STRONG>Instruction fetching is halted</STRONG>. Only those clocks for TAU, time base, decrementer, and JTAG logic remain running. The MPC7451 goes into the doze state to snoop memory operations on the bus and then back to nap using a QREQ/QACK processor-system handshake protocol.

What happens is a hardware interrupt causes the CPU start grabbing instructions again. From an accounting standpoint the idle thread is taking a large number of the CPUs measurable clocks, but it is not actually processing instructions for most of them.

Louis

[ 12-21-2001: Message edited by: lgerbarg ]

[Steve Sell]

Originally posted by Laurence:
<STRONG> idle_thread is 0 and kernel_task is hovering around 90-95%. </STRONG>

I get this same exact thing on my 400MHz PowerBook G4. It's driving me insane. The only way to get it to stop is a restart. Even then it seems to come back. Mostly seems to happen after sleep, but it has just happened at random times. I did not happen before I installed 10.1.2, but I think I upgraded to iTunes 2.03 on the same day.

i hope someone figures this out soon. Funny thing is I have a B&W G3 running on the same config and it doesn't have the issue at all.

-Steve

[jasong]

are any of you running the Palm Desktop beta? When I put my laptop to sleep and wake it up, the kernal task PID 0 goes wild. Uninstalling the beta fixed the problem for me. I had this problem with both the "unofficial" beta (beta76) and the "official" beta (beta 77).

[Ron Goodman]

I saw a similar problem on my wife's Pismo shortly after installing the Palm beta. I haven't heard any complaints about it lately--maybe she trashed the software, since she could never get it to synch.

[Erston]

Hi all,

I'm having the same problem with my 450 Cube [10.1.2, 1.25 GB, SoundSticks, 60 GB ext firewire drive]. I noticed it within the last 10 days, and it seems to be getting worse. I can't notice a pattern to trigger it. Sometimes it occurs when iTunes is running, Explorer, Quake (UGHH!!), mail, whatever. When I can manage to get a top command running, I see kernel_task-PID 0 hogging 80% to 110% of the CPU cycles. I do not have Palm Desktop beta installed, and my iTunes is still v. 2.0.1 (I don't have an iPod yet :-( so I haven't upgraded). I REALLY don't want to reinstall OSX- that's what Windoze people do with their problems There must be something I can do.

Right?

[Erston]

Another factoid that I think is important- I leave my machine on 24/7. I like giving the system the oppurtunity to tidy up with it's early morning cron jobs.

[yukon]

hate to bump an old thread (if it is old-got from search), but this "Kernel_tas) just popped up. I installed the "new AppleSCCSerial.kext-hangfix.sit" for the PPP hang bug, which might be fixed, but suddenly I get the spinny cursor a lot more, and kernel_task is now in top, eating lots of processor cycles.

another thread, http://forums.macnn.com/cgi-bin/ulti...&f=46&t=003664
someone got the error-
************An SCC Error Occurred 0x7C
leading me to bleieve it is the new kext

I'm going to fsck, then if it isn't gone, I'll think about reinstalling OS X again. I'm idly watching my cpu monitor jump up and down from 1 to 12, when it used to hover just around the first block (cpu monitor). switching apps results in the cpu being used totally.

thanks. can anyone explain this kernal_task popping up?

[Detrius]

Originally posted by aenemated:
<STRONG>holy shit ... i've got the same thing. here's from my top ...

0 kernel_tas 1.7% 1:29.88 26 0 - - - 41.5M 558M
0 idle_threa 57.9% 6:47:11

wtf is idle_threa that's eating 57.9% of my CPU?</STRONG>

PID 0 is the kernel.

Idle thread is just what it sounds like. It's a thread that runs when your processor is idling. IE: 57.9% of your processor is NOT being used.