On Tue, 2008-01-08 at 13:55 -0600, Douglas McClendon wrote:
Nils Philippsen wrote:
> On Mon, 2008-01-07 at 21:46 -0600, Douglas McClendon wrote:
>> Jeff Spaleta wrote:
>>> On Jan 7, 2008 4:28 PM, Douglas McClendon
>>> <dmc.fedora(a)filteredperception.org> wrote:
>>>> I had been happily just yanking usbsticks for a long time, though just
>>>> the other day I did manage to corrupt a vfat fs on one pretty badly.
>>> But would you ever want to yank a terabyte sized usb "stick"
like
>>> Maxtor OneTouch III Turbo 1TB External Drive?
>> Absolutely, if I trusted that the kernel code was smart enough to never
>> result in a corrupted filesystem.
>
> Hm. I had the impression that even Windows nowadays has icons to
> "unmount" removable media (like USB sticks), possibly for a reason?
why I referenced win3.1
But then win3.1 is not the shining example of robustness we should
thrive to emulate, now is it ;-).
> Anyway, you'd also have to trust that the hardware can cope
with being
> yanked, possibly in mid-operation.
Doesn't sound like an impossible engineering challenge to me. Also why
I referenced waiting 3X the time you know that the OS config would allow
writes to be in flight (or ye ol sync mount option).
Well, not theoretically impossible, but if I recall correctly there are
IDE/ATA/SATA(?) -- eeh -- consumer drives out there that don't even
guarantee bits to be on disk when they should be. I think Alan wrote
something about that a while ago.
Nils
--
Nils Philippsen / Red Hat / nphilipp(a)redhat.com
"Those who would give up Essential Liberty to purchase a little Temporary
Safety, deserve neither Liberty nor Safety." -- B. Franklin, 1759
PGP fingerprint: C4A8 9474 5C4C ADE3 2B8F 656D 47D8 9B65 6951 3011