Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Turning Off the Computer

Ward Vandewege wrote:
> On Sat, Oct 18, 2008 at 09:32:17PM -0400, Jarod Wilson wrote:
>> On Sat, 2008-10-18 at 21:13 -0400, Bruce Borland wrote:
>>> The other night I asked my son to turn off the computer.  He apparently 
>>> did not want to wait for the computer to shut down normally, so he 
>>> decided to switch off the power to the machine, stopping it immediately. 
>>>  I told him that was not good to do, but he asked me why.  I did not 
>>> have an answer for him.  He did this to our Windows machine, but I 
>>> understand that it is not good to turn off a Linux machine this way 
>>> either.  Can someone explain what problems are caused by such an 
>>> immediate shutdown?  I would like to know, and would like to tell my 
>>> son, too.  Thanks.
>> Hard disks are slow, so operating systems tend to cache data into
>> volatile system memory, before periodically flushing it out to disk --
>> only doing one larger batch write vs. lots of small writes all over the
>> place is a performance win. Pulling the machine in mid-flush means the
>> on-disk data winds up in an inconsistent state. In some cases, a hard
>> drive that looses power while in the middle of writing data can even
>> cause physical damage to the disk platters (rare anymore these days
>> though). 
> That was what the 'disk parking' feature was for. Remember, a long long time
> ago, that a drive had to be 'parked' manually before you could cut power to
> it? 
> Drives have been 'auto-parking' for a long time now (cf
> I
> certainly have not seen a drive in the past 15 years or so that needed manual
> parking.
Right. Parking the head is not the issue - making sure all the disk 
buffers are flushed is the issue, on both Windows and Linux. Both 
systems would be drastically slower if they didn't cache. Both systems 
also automatically try to recover - as best they can - from bad 
shutdowns, but there's always a risk of losing data and, if nothing 
else, wasting time cleaning up the mess.


> Thanks,
> Ward.

BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!

Boston Linux & Unix /