I had a strange problem appear after I spilled tea on my 2010 MBP last year. Continually, kernel_task was using an entire core (150-250% in top) and no amount of reboots could fix this. The built-in hardware diagnostic indicated it was not compatible with my Mac, and I could find nothing else that apparently was wrong. My computer was running very slowly though, and the system load was always over 4.
Finally I found through research that this is somehow related to power management. Apparently the system will have the CPU loop simple operations to prevent the processor from doing actual work and heating up. Sure enough, disabling power management by hacking a kernel extension made the kernel_task problem go away. My Mac has run perfectly for three months ever since I made that fix, after being nearly disabled for a year.
I remember having this same problem on a hackintosh. Installing the NullPowerManagement kext was the usual fix back then since DSDT patches were brand new and not well known by the community. If that kext is still available and compatible with 10.9 it's probably a good way to handle this until they fix it.
Had the exact same problem (minus the spilled tea, but I'm often running my computer very hot so I'm not surprised) on OS 10.8 on a 2011 (?) MBP, and did exactly this to fix it. I verified that the problem was with some sort of IntelACPI... [sic] module using a DTrace script (been a while but I can try to find it if anyone wants... basically that module was using up a large amount of CPU whenever there was load, such as when I launched a game. Long story short, the step-by-step to fix it is here:
Been working like a charm for about 13 months now, and it's survived multiple minor 10.8.x upgrades since then. YMMV may vary with 10.9, and since they supposedly changed power management, I wouldn't touch that upgrade with a [Mac OS] 10 foot pole.
The fix for me was to simply delete the power profile from my MacBook model in a kext config directory. Everything was fine until I upgraded to 10.8.5, and then to my system would not boot at all – feeling with the kernel panic very early in booth that had something to do with AppleIntelCpuPowerManagement.kext.
I completely reinstalled and install Mavericks, and then repeated the original fix. It worked fine and my system is doing great running Mavericks.
Uh, "feeling with the kernel panic very early in booth" meaning, boot failed with a kernel panic.
I was kind of confused by what I read about power management having been changed in Mavericks, but it turned out that whatever is relevant to this issue has not changed. As I noted the same solution fixed the same problem.
Damn, my 2009 Macbook has been running very slowly for the past 4 months or so and no amount of rebooting would resolve the problem. I was actually puzzled when it slowed down to a crawl becoming basically useless and would even restart with the same slowness. Could you elaborate on your fix?
Also try running sudo fs_usage in Terminal. Every line printed is a disk read/write and since you would be on an older machine with slower I/O you need to reduce the number of lines printed.
One useful trick is to add folders to Spotlight's Privacy section e.g. source code, downloads, torrent locations, game files etc.
Any other company that would DARE completely ignore the cries for help of its consumers would be publicly shamed and their reputation would suffer.
Yet, not a single word in this thread about how silent they tend to be about issues effecting thousands of users (which is a common trend with them)...
I wonder how they manage to pull that off, it's magic.
Really? Name one non open-source company that will jump on forum threads and fix it, rather than just putting it in their internal bug tracker to investigate.
I have a Logitech gaming keyboard. Their drivers are shockingly bad and freeze up every so often requiring me to re-plug the keyboard to get them to to work again. This is mentioned in the forums for at least 6 months, no fix has been announced.
I'm not defending Apple here, they should at least comment on their own forums, but they're not alone.
Back in the day, when there was no internet (only the arpanet) and Unix systems shipped with 10 manuals in nicely bound cases (I had one I was in my 20's at the time) things moved much slower and a vendor would never consider acting like this because they served primarily business customers (and institutions). Just the memory on my first Unix system was like $4000 a MB. It came with a 70mb hard disk.
To me the change most likely came around 1981 when the IBM PC came out and was sold to consumers who basically believed that it was them that was stupid and not something wrong with the machine. Manufacturers then learned that they could shove things into the pipeline and fix them later and that they actually got a fair amount of slack in the marketplace from the "stupid" end users. People who knew their shit using these machines were totally outnumbered by computer illiterate people. In other cases the ubiquitous "tech guy" made money because things didn't work so who was he to complain? That's what kept Microsoft in power. The tech guy ecosystem.
I'm curious if anyone else agrees at all with my take. Strictly my opinion having been around computers so many years.
If you have 1 million $10 customers and 10,000 have a problem the cost of losing their business isn't amazingly huge, especially as most of them probably won't switch vendors.
When you have ten $1,000,000 customers you might listen more if one has a problem.
Yet you can still do this and hang Visual Studio 2012, and maybe 2013:
* On a new machine (or maybe a new profile, I haven't checked), enable PowerShell's AllSigned option
* Open Visual Studio and open an MVC4 project, bonus points if it's close to the template in terms of installed NuGet packages
* Watch as Visual Studio hangs indefinitely.
This has been reported a number of times to Microsoft, but I don't think it was fixed in Update 4.
What's happening is PowerShell, running inside Visual Studio is prompting for something to do with a signed PowerShell Script, but Visual Studio hasn't loaded enough of the UI to let you accept the warning/prompt.
That could be because Apple's forums are not the venue for support. For developer support, there's rdar. For customer support, there's AppleCare. They state this pretty clearly, and there are perennial reminders when the iOS betas come out and the forums get flooded with non-devs who brick their iPhones.
Apple gets my repeat business (read: magic) by solving my hardware and software problems reliably and quick when I present them through the appropriate channels, go well over a decade now. For one particular repair, I was given a loaner machine, and another resulted in being updated to newer hardware - common stories amongst people who use their AppleCare.
Simple - a few thousand (if it is that many) people out of several million people having a problem is not always evidence of a real problem but rather proof that with a large enough install base some people will always have problems and it might well be for no reason at all beyond shear chance.
If there are 10 million installations of Mavericks (and I've got no idea how many there really are) and 5,000 people reporting a particular problem, that's only .05% of users - which is well within the expected proportion of users you'd expect to have problems with any upgrades. Note, for comparison the recent PS4 release problems and Sony's statement about the 0.4% of users experiencing problems - "This is within our expectations for a new product introduction"
Except you have no idea how many people are affected, you only know how many A) noticed, B) traced the culprit back to kernel_task, and C) reported it. That's likely a very small percentage of the afflicted population.
I'm also not sure how you explain a CPU pegged by a kernel process on a machine which should be doing almost nothing as "shear chance". That's a bug.
>Any other company that would DARE completely ignore the cries for help of its consumers would be publicly shamed and their reputation would suffer.
Completely ignore? Like several public statements, recall programs, warranty extensions and other such stuff they have done for problems in the past?
They have the bug in their radar bug tracking system, they close duplicate reports (which means they know about it specifically), and they are obviously working on solving it. What can they say before a fix is ready? "We're working on it"? Of course they are.
Perhaps you missed the fact that they top the user satisfaction ratings almost every year.
I recently picked up an old MBP and the logic board die on me. Turns out the faulty GPU unit was a known issue, but they never did a recall. Instead they extended the warranty portion for this part by four years.
This is the first real issue aside from iPhone battery woes a few years back.
The mmore I look through forums and such, the more I find Apple not shipping products that meet the market perception of a superiority.
I haven't been able to find one word from them about how they broke everyone's VPNs with Mavericks. It's fine though, I just consider it our just punishment for ever allowing Macs into a business environment.
Can you say a bit more about that? I have some 10.8.x machines that I was going to upgrade to 10.9.x in order to fix some VPN problems that I am having with that. What's up with Mavericks and VPN? (Links, commentary etc..)
Mavericks (and server 3.0) was released on October 22nd.
December 19th the update was released for their vpn issue. Less than two months does not equal "like half a year".
The issue was vpn behind nat wasn't working as expected, copying raccoon over from 10.8 server fixed the issue (not ideal but it took all of 5 mins).
The so called Mac premium pricing (which isn't really true if you compare like for like) is for the superior hardware. Better trackpads, cases that don't flex, better keyboards, better overall design. The issue here is likely a software bug for an OS that was only recently released.
You would think someone on HN would know the difference.
Worst cables ever... so far for the 'superior' hardware.
Prices are at their fairest the moment they release new hardware. The later you buy their hardware the bigger the difference with their competitors (They only lower the price of older hardware when releasing a new model and not that much either).
But superior hardware, no, that is a bridge too far. Decent hardware, yes, but not all hardware from Apple is equal.
I've had this problem with all my Macs at home, although it would never occur with other Wi-Fi access points. This leads me to believe it has something to do with the wireless access point itself.
I was never able to fix the problem, but it went away after I bought an AirPort Time Capsule, which now serves as my wireless router. It has an extremely limited feature set compared to a real router, but I haven't had connection problems once since implementing it.
Do you use a bluetooth keyboard? My wifi flakiness seems to be much less of a problem when my lid is open and I'm typing on the builtin keyboard than when the lid is closed and I'm using the Apple bluetooth keyboard.
It's so intermittent that it's very hard to diagnose, but that's one thing I've noticed.
The wifi problem drives me absolutely nuts. Most of the time I'm somewhere I can use a wired connection but the lack of a solution is incredibly frustrating.
In my opinion, Mavericks is the worst OS X version so far. It's caused a handful of crashes/hangs to my Macbook Pro, which I only encountered with about once or twice a year with Snow Leopard, Lion and Mountain Lion. I'm not sure it is caused by the OS itself or the compatibility of the apps I installed, but it's still bad.
I just had a crash few mins ago after I reinstalled the OS entirely last week. According to the crash log, it has something to do with the "kernel_task". That sucks!
As someone thats been on the osx train since 10.2. I get somewhat amused by the "OSX N+latest is the worst OSX" comments.
I had the (dis)pleasure of using someone's 10.1 machine, oy wow, surprised that even got released.
As a counterpart, mavericks has been the most stable osx for myself so far. Additionally the battery life improvements gave me an extra hour on battery alone. You can pry mavericks from my cold dead hands at this point. The energy tab is also really useful at finding out what is using up battery needlessly (looking at you chrome/firefox).
I feel jealous with your luck. I've had as many crash as I can imagine since switching to Mac from Windows (before, I don't even remember how many times I got freezing apps or BSOD). And it seems to occur that often since Mavericks so it's my sole assumption.
I agree about the battery life improvements part as a nomadic user. However, it's totally depends on which applications you are using. For me, chrome/firefox, which eats up most of the battery, is used constantly, therefore, cannot be turned off (to save energy, because it doesn't support AppNap, I suppose). I don't see any help looking at the Energy Tab as whatever it's like, I don't have a choice there.
Each major release of OS X, I know that I should wait until the x.1 release, but I get excited and install the x.0 release anyway. And every release, the same thing happens. Tons of bugs, crashes, and frustration and I swear that next release I'll wait until x.1.
Problems I have, on a brand new 15 retina MBPro:
1- after wake from sleep, wifi needs to be turned off and on to find anything
2- no sound from internal or external speaker until restart, although the OS behaves as nothing is wrong, the onscreen indicators and system preferences act as if they are muting and unmuting and changing the volume
3- kernel panics every now and then
4- display jitters (like graphic buffer corruption) a couple of times
5- after wake up from sleep , keyboard does not respond while track pad is, which means I cannot login, needs force shutdown,
If Apple cannot solve your problems, within the terms if your warranty, they will offer alternatives (eventually) - up to and including a refund.
I understand people want to raise issues to "public awareness" that they feel are something Apple should pay attention to. And in the case of the original submission, VERY SPECIFIC details were figured out - and Bug Reported to Apple! Apple even closed some of the reports as a duplicate. They know about this problem - and it's for a new OS that has only been out for a handful of months - and the issue is for specific hardware builds.
They are working on it.
You, however, have a laundry list of real issues - USE YOUR WARRANTY.
thanks to all people who commented, I don't want my money back , I want a working MacBook, and it is not easy to bring this up to and apple store, since it does not happen all the time, I need to go there and try to reproduce it, and they I expect they say let's try a OS reinstall, then I'll go home and then it is likely that it does not solve some of the problems and ....
it is not what I expect from what I consider best laptop (my) money can buy,
Some of those (minor) problems, in fact most, are the standard kind of "new OS" bugs. None of them sounds like its about hardware faults. So your advice is not really applicable here. He'd better wait for 10.9.2 and such.
My advice is 100% applicable. Especially with issue #5, but in general as well.
You don't seem to understand: He is unhappy with what he purchased because he is experiencing problems that should not exist in the product. A brand new product. And not just a single issue - but several of them!
This isn't a device that (in his personal experience) worked fine at some point in the past on a different version of software but now is having problems on the latest software update.
This is a device that, brand new from Apple, has problems. This point is especially relevant as you cannot downgrade a Mac from Apple to an OS version that came out prior to the introduction of that specific model of machine. The Late 2013 MacBook Pro Retina devices shipped with Mavericks. They will never run Mountain Lion, you cannot avoid Mavericks on them for the time being.
This is the entire point of a warranty (and lemon laws in many locations): You don't have to "wait" for something you just bought to get better / work properly.
You take it back and show them the problems you're having.
Either they can fix it or they can't. If they can't fix it, he can get his money back or some other alternative offered by Apple that he might agree to (maybe a different model, or a complete replacement).
His money is not trapped in this device, so he doesn't have to put up with this experience - unless he actually prefers to complain and be unhappy.
>You don't seem to understand: He is unhappy with what he purchased because he is experiencing problems that should not exist in the product. A brand new product. And not just a single issue - but several of them!
Which product came with a brand new OS version. And which, like most OS launches has several software issues. Which are not gonna be solved if he returns his device (or only randomly, if they affect just one of a few Apple OEM partners chipsets etc).
>This is the entire point of a warranty (and lemon laws in many locations): You don't have to "wait" for something you just bought to get better / work properly.
That only holds if it's a faulty device. For OS and driver bugs, you very much have to wait. And no matter how long you wait, there will always be some bugs to, err, bug you, in it.
I must strenuously disagree. If these were software problems then online forums and Apple tech sites would be filled with folks tearing down the virtual walls. If you get a box with bizarre stuff going on and a search doesn't turn up a screaming horde with exactly your symptoms, then in all likelihood it's a local problem. That is a silver-lining to the decreasing post-purchase configurability of the Macbook Pro line: it makes it a heck of a lot easier to diagnose potential hardware issues.
Taken all together, these almost certainly point to bad hardware. It's possible that a bad third-party kernel extension might be the culprit as well, so testing in SafeBoot can be useful to ferret out such issues.
Last but not least, I'm personally making heavy use of the same hardware and OS daily with zero issues. I'm pretty sure it's not the Apple-supplied software.
>I must strenuously disagree. If these were software problems then online forums and Apple tech sites would be filled with folks tearing down the virtual walls. If you get a box with bizarre stuff going on and a search doesn't turn up a screaming horde with exactly your symptoms, then in all likelihood it's a local problem.
That a bug is a software bug doesn't mean that it manifests in all identical systems. It depends on various factors (ie. combination of third party peripherals, e.g some Macs have Samsung SSDs, others have other brand), install software, user workflow, etc.
>Why does it matter whether it's a hardware or software problem? If it's broken it's broken.
Of course if it's a software problem he shouldn't expect a hardware replacement to fix it, but if he wants a refund he should get it.
It matters because software problems will always exist, there is no bug free software or OS, and the important ones get fixed in upcoming upgrades.
What would he do with his refund if it was a software problem? Switch to some magical unicorn non-issue OS or get another identical machine and OS?
> 5- after wake up from sleep , keyboard does not respond while track pad is, which means I cannot login, needs force shutdown
I have a similar problem sometimes and it appears to be some kind of USB bus / driver issue. If you plug in a USB keyboard, wait for it to be recognized, then unplug it the built in keyboard almost always starts working again. I've seen USB related error messages show up in logs when this happens.
Before you put the laptop to sleep unplug all USB devices and make sure the keyboard and trackpad are functioning correctly.
This must be somehow hardware related. I went through 5 rMBPs, because I got bad screens 5 times in a row. There is an entire 600 page thread on that as well. That being said, none of the 5 had this problem. Does it do the same thing with a clean install of OSX?
You can clone your current installation to a USB hard drive and install OSX from scratch. Carbon Copy works well, or just use dd. If it does the same with a clean install, take it to Apple, and make them deal with it.
Holy cow, get warranty service ASAP. As someone with a brand new 15" rMBP, I do not ever have any of these problems. My system is absolutely rock solid in heavy daily use.
My Mac freezes a few times a week. Just the window manager, because I can still ssh in. I know it isn't the lid, because it is an iMac, but it may be going to sleep that kills it.
i posted a real fix below. sadly this fix the one you mentioned doesn't always work for me. also, usually i wouldn't notice that the bug happened until my battery was half drained
Another (much more) common issue which seems to be related to Mavericks is the Finder problem where clicking Finder results a sudden hang of the system a the spinning colour wheel.
I'm fairly new to Apple products (~2 years ago) this is the worst problem so far but I'm surprised at such little feedback from Apple on it.
Not Mavericks-specific, but Finder is notorious for hanging when mounted volumes become unavailable or unreliable. And it's certainly possible that some of the network filesystem changes in Mavericks — in particular, it now uses SMB in certain cases where previous versions would have defaulted to AFP — could lead to this behavior in situations where you didn't see it before.
Solution on a late 2013 MBP for me (when the issue occurs no sound is playing anymore - no system sounds, no videos, no audio): Plugging in earphones and unplugging. Killing the coreaudio process as recommended at a few places online doesn't do the trick, unfortunately.
I've been wrestling this as well, and found this post too. Another problem I frequently wrestle with is my air sleeping on not coming back (the keyboard stays lit though). Only way to get it back is too reboot.
I experienced this right out of the gate after upgrading. I was very curious to see what my battery savings would be and decided to ditch my charger, only to have my battery gone in an hour. It's since subsided– I'm not really sure why, though at one point I did about everything I could think of short of reinstalling. I removed all kernel extensions I had added (KR4MB for example), removed SIMBL, reset PRAM, cycled the battery, etc. Now the only issue I have is about 10s after closing the lid, the fan goes nuts for another 10s before going to sleep.
My personal list of gripes (1st-gen 15" MBPr) is short, but it's sad that they all appeared after the Mavericks upgrade:
1. wifi will fail to reconnect on wake.
2. almost 40% of RAM (about 6 GB of 16) in use right after boot with no apps running. Was about 15% before Mavericks.
3. I got the dreadful Keychain error once, luckily managed to fix it (I use FileVault, so no SafeBoot for me and losing the key would have been fairly disastrous).
All these very clearly appeared after I took the plunge and moved to 10.9.1. Somebody screwed up.
It's related to the audio driver and is triggered most frequently when you close the laptop with headphones in. Simply unplug them, close the laptop, and wait for 5-10 seconds. reopen... you will still see the cpu spike, but it will end within a few seconds of everything awaking... it is annoying, but this seems to resolve it for now.
I gave up on Mavericks after it went on an irreparable rampage of asking me for a "keychain password", then after trying a fix it just wouldn't boot anymore.
It required me to recover using Time Machine (and had a lot of workarounds, because it can't recover booting from the recovery partition in the HD, sigh)
If you boot into recovery drive and go under Utilities -> Terminal type 'resetpassword' and in that same window there will be a ACL ( OS X Access Control Lists ) reset for an account. If you do that when you reboot, it prompts you to make a new keychain before logging you in.
I've been running into this problem with my 2013 MBA. It happens on wake, and the only that fixes it is a reboot. I'm getting pretty fed up with the issue since it's tanking my battery life and there seems like nothing I can do to fix it permanently.
Unrelated to this particular issue, but lately Adobe Updater has taken to running lsof in a tight loop, waiting for me to exit Acrobat so it can update. Makes for a pretty solid red bar of system CPU use.
Could also be other issues unless you see this specific set of symptoms. Mine gets sluggish but it seems to have something to do with the 'systemstats' process. When I open up something that queries it (e.g. click on the power icon in the menu bar to make it go to "Collecting power usage information...", or open Activity Monitor), it goes nuts, pegging the CPU and ballooning to >2GB of memory. Haven't figured out what to do about that, besides kill -9'ing it and avoiding doing anything that'll cause it to respawn.
Moving the /private/var/db/systemstats directory out of the way and rebooting seems to have solved the problem. Guess something there was corrupt or too large.
I noticed slowdown on my brand new Air with nothing installed (beside a browser, xcode, and VLC). I have to reboot multiple times per day, closing the lid doesn't help, it causes it. Really unacceptable and a lot of people are seeing the same thing.
I'll be regretting not going with a Lenovo until they fix this.
Antennagate? You mean that old non-issue blown up by some media out of all proportion, that happened to phones before and since (including competitors phones), and despite which "issue" the iPhone 4 went on to sell further tens of millions of devices with the same design?
Finally I found through research that this is somehow related to power management. Apparently the system will have the CPU loop simple operations to prevent the processor from doing actual work and heating up. Sure enough, disabling power management by hacking a kernel extension made the kernel_task problem go away. My Mac has run perfectly for three months ever since I made that fix, after being nearly disabled for a year.