1

My MBR start freezing up for over hour and I've found that my kernel reported being throttled, e.g. (this is just a sample for 2 seconds)

$ sudo fs_usage | | grep -i throttled
12:26:45    THROTTLED    0.015159   kernel_task 
12:26:45    THROTTLED    0.015606   kernel_task 
12:26:45    THROTTLED    0.016328   kernel_task 
12:26:45    THROTTLED    0.014862   kernel_task 
12:26:45    THROTTLED    0.014797   kernel_task 
12:26:45    THROTTLED    0.015583   kernel_task 
12:26:45    THROTTLED    0.014794   kernel_task 
12:26:45    THROTTLED    0.014668   kernel_task 
12:26:45    THROTTLED    0.014915   kernel_task 
12:26:45    THROTTLED    0.014555   kernel_task 
12:26:45    THROTTLED    0.014660   kernel_task 
12:26:45    THROTTLED    0.015547   kernel_task 
12:26:45    THROTTLED    0.015364   kernel_task 
12:26:45    THROTTLED    0.015073   kernel_task 
12:26:45    THROTTLED    0.014832   kernel_task 
12:26:46    THROTTLED    0.015649   kernel_task 
12:26:46    THROTTLED    0.016643   kernel_task 
12:26:46    THROTTLED    0.014496   kernel_task 
12:26:46    THROTTLED    0.015488   kernel_task 
12:26:46    THROTTLED    0.014576   kernel_task 
12:26:46    THROTTLED    0.015531   kernel_task 
12:26:46    THROTTLED    0.015224   kernel_task 
12:26:46    THROTTLED    0.015182   kernel_task 
12:26:46    THROTTLED    0.015443   kernel_task 
12:26:46    THROTTLED    0.015166   kernel_task 
12:26:46    THROTTLED    0.015435   kernel_task 
12:26:46    THROTTLED    0.015105   kernel_task 
12:26:46    THROTTLED    0.015091   kernel_task 
12:26:46    THROTTLED    0.015131   kernel_task 
12:26:46    THROTTLED    0.015259   kernel_task 
12:26:46    THROTTLED    0.014865   kernel_task 
12:26:46    THROTTLED    0.015681   kernel_task 
12:26:46    THROTTLED    0.014455   kernel_task 
12:26:46    THROTTLED    0.015251   kernel_task 
12:26:46    THROTTLED    0.015928   kernel_task 
12:26:46    THROTTLED    0.015053   kernel_task 
12:26:46    THROTTLED    0.014699   kernel_task 
12:26:46    THROTTLED    0.015251   kernel_task 
12:26:46    THROTTLED    0.014499   kernel_task 
12:26:46    THROTTLED    0.015589   kernel_task 
12:26:46    THROTTLED    0.015166   kernel_task 
12:26:46    THROTTLED    0.015422   kernel_task 
12:26:46    THROTTLED    0.018772   nsurlsession
12:26:46    THROTTLED    0.029125   kernel_task 
12:26:46    THROTTLED    0.014671   kernel_task 
12:26:46    THROTTLED    0.014763   kernel_task 
12:26:46    THROTTLED    0.015114   kernel_task 
12:26:46    THROTTLED    0.009341   iStat Menus 
12:26:46    THROTTLED    0.008501   kernel_task 
12:26:46    THROTTLED    0.014090   kernel_task 

I've increased fan control to the maximum (via iStat Menus), as I've read it may be related to the high CPU temperature, but it didn't do any difference. Only killing some tasks helped a bit.

I know how to kill the tasks or restart the computer to temporary fix the problem. However I'm more interested what exactly is the kernel throttling, why this happens (why it's throttled and what's preventing it from not being throttled) and what action may I take or how to unthrottle it? Maybe there is some option to make it less strict or disable it, or maybe decreasing CPU temperature further more may help?


Additional details:

$ uname -a
Darwin MBP.local 15.2.0 Darwin Kernel Version 15.2.0: Fri Nov 13 19:56:56 PST 2015; root:xnu-3248.20.55~2/RELEASE_X86_64 x86_64 i386 MacBookPro11,3 Darwin
$ top | head
Processes: 464 total, 19 running, 8 stuck, 437 sleeping, 4157 threads 
2016/05/25 12:31:44
Load Avg: 30.72, 39.86, 29.86 
CPU usage: 14.86% user, 46.0% sys, 39.12% idle 
SharedLibs: 84M resident, 18M data, 7760K linkedit.
MemRegions: 360850 total, 2901M resident, 55M private, 435M shared.
PhysMem: 16G used (8494M wired), 5664K unused.
VM: 1406G vsize, 528M framework vsize, 3900541086(0) swapins, 3913990343(0) swapouts.
Networks: packets: 136234033/128G in, 133429405/73G out.
Disks: 288046480/15T read, 220716626/15T written.
$ macosver -h | tail -n1
ProductVersion 10.11.2
1

fs_usage shows information about file system usage (and related things). Given that, I suspect THROTTLED from fs_usage indicates it is disk access being throttled and not CPU (although THROTTLED is not clearly documented in the man page). OSX reduces I/O throughput of idle applications to give better performance to active ones (see this Ars article) so it's not necessarily a problem.

I think the important thing here though is that this isn't evidence, one way or another, your CPU has been throttled due to heat and that that is causing your hangs.

Might be worth looking and disk I/O however. In most traditional UNIXes low disk through put can cause apparently hangs - maybe that what you are seeing - but you would need to investigate more - or we can only speculate.

|improve this answer|||||
  • Top shows Disks: 52T read, 52T written., I assume it's since my 37 day uptime, so it's like 1.4TB each day = ~0.5 peta bytes each year, so after 3 years I've probably reached over 1.5PB of written data (MBP Late 2013), so you're saying there is a chance that my SSD may be dead? – kenorb Jun 21 '16 at 18:44
  • No I'm not saying that at all. OSX throttles I/O of idle applications. This is normal behaviour. Whilst it isn't documented I believe the output of fs_usage is just showing you that behaviour. Therefore what you have pasted above does not give us any great insight into why your machine is freezing. My comment about the amount of disk I/O going on was just something to check, given we so far have little idea what might be happening. You would need to look at this value whilst your machine was freezing. – user133831 Jun 21 '16 at 18:51
  • 1
    Note also that your machine may be throttling the CPU due to heat - I just don't think that the fs_usage output shows that one way or other. – user133831 Jun 21 '16 at 18:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.