Re: crash after using flash drive

E Z Peaces wrote:
The Natural Philosopher wrote:
E Z Peaces wrote:
Fred McKenzie wrote:
In article <f78nff$pur$1@xxxxxxxxxxxxxxxxxxxxxxxxx>,
E Z Peaces <cash@xxxxxxxxxxxxxxx> wrote:

Yesterday I woke my Mac (10.4.10) and checked USENET. In a minute it said I had disconnected a USB drive without ejecting it. A minute later the screen went dim and a sign told me in four languages to shut down.

I suppose that was one of those Colonel Panics I've heard about. An hour ago, just before putting the computer to sleep, I had disconnected a flash drive, and perhaps I'd forgotten to eject it first.

Could that have caused a panic a couple of minutes after the computer awoke?

E.Z.-

The one kernel panic I've had filled the screen with garbage text and locked-up the computer. It was caused by a broken ground wire in a USB mouse. I had no prior warning or opportunity to shut down. Your experience may not have been a true kernel panic as I understand it.

USB is a "hot plug-able/unplug-able" system. In theory there is no electrical damage done when connecting or disconnecting while power is applied.

Normally the message telling about an improperly disconnected flash drive is just a warning. The damage that might occur would be if the flash drive was being written-to when disconnected, and the damage would be corrupted data and/or format on the flash drive.

Fred

My panic log called it an unresolved kernel trap. Is that a panic?

Sort of. It means the kernel probably did something outside of anything it should have, and being as there is nothing more basic than the kernel to deal with it, instead of being shutdown, it shut itself down.

I have in the past unplugged USB storage without remembering to eject it.

You run the risk of an inconsistent file system if you do that.

The difference this time was that I put the computer to sleep
before it warned me.

I think its a bug worth reporting to the developers. I created a bug like this once..failing to close files when errors were encountered..clumsy operators eventually used up file handles causing a crash, when a file open on a valid file failed.. You have obviously found a way to evoke a condition the programmers haven't considered.

Not sure how one reports kernel crashes to the apple boys..


I think it filed a crash report automatically.



I had a complete system fail, that would not repair by ANY of the standard battery of methods, including archive/install on a new 3 day old) iMac 24 late last year. Even Apple were happy to replace it as a DOA. Only resort to full wipe and replace the system brought it back.

The culprit: a defective USB thumb drive I had been using for a couple of years. Yes It was the drive, I reproduced the crash/wipe/reinstall as confirmation.


regards

Chris
.

Relevant Pages

  • Re: crash after using flash drive
    ... The one kernel panic I've had filled the screen with garbage text and locked-up the computer. ... It was caused by a broken ground wire in a USB mouse. ... In theory there is no electrical damage done when connecting or disconnecting while power is applied. ... I think its a bug worth reporting to the developers. ...
    (comp.sys.mac.system)
  • Re: crash after using flash drive
    ... E Z Peaces wrote: ... The one kernel panic I've had filled the screen with garbage text and locked-up the computer. ... It was caused by a broken ground wire in a USB mouse. ... In theory there is no electrical damage done when connecting or disconnecting while power is applied. ...
    (comp.sys.mac.system)
  • [2.6.6] oops in khubd on usb bluetooth disconnect
    ... after shortly using a usb bluetooth adapter ... with rfcomm to a Nokia 6310i disconnecting the adapter caused this oops ... Unable to handle kernel NULL pointer dereference at virtual address ...
    (Linux-Kernel)
  • PROBLEM: Oops when using emi62 with 2.6.28
    ... With kernel 2.6.28 (and I don't really know which version of the kernel ... emi62 soundcard usb firmware oops ... 2000-2fff: PCI Bus 0000:02 ... # Infrared-port device drivers ...
    (Linux-Kernel)
  • problem with linux-2.6.28
    ... ACPI: Local APIC address 0xfee00000 ... Allocating PCI resources starting at cc000000 ... 3243604k/3276480k available (3222k kernel code, 31608k reserved, 1142k data, 312k init, 23 ... USB 2.0 'Enhanced' Host Controller Driver ...
    (Linux-Kernel)
Loading