[NTLUG:Discuss] Is there a disk defragmenter for Linux

Mon Dec 23 00:16:13 CST 2002

Vaidya, Harshal (Cognizant) wrote:

>I don't have any idea about ext2 but ext3 which is a journaling file system
>boasts no need of fsck too. Although they have included fsck for ext3 they
>say that you will hardly need it. Moreover, as the FS is journalling it is
>completely free of fragmentations in normal web server or data base
>environments with conventional loads. 
>
>Although I haven't digged in deeper into the technical details of why? and
>How? I am just sharing the information which I had once read on
>Linuxgazette.com
>
>
>Regards,
>Harshal Vaidya.
>

I'd like to see that article, if you happen to know where it is.

>
>-----Original Message-----
>From: terry [mailto:linux at cowtown.net]
>Sent: Sunday, December 22, 2002 11:20 PM
>To: discuss at ntlug.org
>Subject: Re: [NTLUG:Discuss] Is there a disk defragmenter for Linux
>
>
>The short answer is yes.
>
>Steve Baker wrote:
>
>  
>
>>Rev. wRy wrote:
>>
>>    
>>
>>>>No - only crappy OS's need to have their disks defragged.
>>>>        
>>>>
>>>Can you expand on this for a moment?  It's a critical difference 
>>>between M$ and Linux, and I've yet to stumble across anything other 
>>>than one line answers
>>>to the question that say any more than what you've written.  Yes, 
>>>I've looked
>>>at the LDP howto on filesystems.  I've googled (although perhaps on 
>>>the wrong
>>>subjects).  I've taken various classes on *nix, but I've yet to see this
>>>explained.
>>>      
>>>
>I've researched this issue myself and have found that although it's 
>probably not a very popular or active project but a disk defragmenter is 
>available for Linux. (You'll find it on fresheat.net if your 
>interested.)   I think however, that it's a trivial pursuit on either 
>platform, but much more trivial on a Linux file system because of the 
>fact that it's just a more organized and efficient system and structure, 
>especially when you have multiple partitions.  (It's probably one good 
>reason to have separate partitions or drives for / and /boot and /var 
>and /usr and /home etc.)
>
>Under normal circumstances, fragmentation just doesn't occur, at least 
>not so much that it matters.  Only if you regularly delete large blocks 
>of data will it become an issue at all, and even then, with the modern 
>hardware we have now, the seek times are so fast that I doubt that you'd 
>notice any performance loss even if you did have a considerable amount 
>of fragmentation.  And the term "large blocks of data" is a relative 
>term. Consider the sizes of hard drives now days, some of us have 1 or 2 
>gig hard drives, others of us have 6 or 10 gig hard drives and still 
>others of us now have 80 or 100 gig drives.  So how much fragmentation 
>would make a difference on a 100 gig hard drive? How much fragmemtation 
>would make a difference on a 10 gig drive?
>
>  
>
>>Well, I confess that I don't understand the technical details - but as 
>>an observed
>>fact, when you run fsck and it reports your fragmentation, it's never 
>>very far
>>off 90% perfect.
>>
>>    
>>
>>>So how exactly does *nix write data to a hard drive that eliminates 
>>>the need for a defrag?  And why is there fsck if there is no need for 
>>>a defrag?
>>>      
>>>
>It writes data concurrently or consecutively.  
>
> I think that if fragmentation were enough of an issue, the various 
>Linux  distributors would include a defrag utility in the default 
>install, but it's not, so they don't.
>
>  
>
>>fsck doesn't have anything to do with fragmentation - it checks the 
>>file system
>>for consistancy - for errors - and tries to repair them.
>>
>>In theory, there should be no filesystem errors - but if your machine 
>>crashes
>>or you have hardware problems - then occasionally, something gets 
>>screwed up and
>>on your next reboot, fsck repairs it.
>>
>>Nothing to do with fragmentation.
>>
>>    
>>
>>>Obviously Linux requires a different way of thinking than does M$, 
>>>but I don't see how saying "Crap OS'es need a defrag, Linux doesn't" 
>>>explains what's going on under the
>>>hood, and often times with Linux, knowing what goes on under the hood 
>>>is half the battle
>>>won.
>>>      
>>>
>>Well, yes - but it's hard to be an expert about everything.  Some 
>>things outside
>>your own field, you just have to take on trust...or if it really bugs 
>>you - become
>>an expert.  After all, you can always read the source code for the 
>>file system and
>>see what clever thing it does.
>>
>>Personally, I'm happy to take it on trust that fragmentation isn't an 
>>issue.
>>
>>Practical experience says it's not.  Many Linux systems run for a year 
>>or more
>>without being rebooted - much less defragged - and nobody notices any 
>>drop in
>>performance in the way you do on a horribly fragmented Windoze system.
>>
>>As you use Linux longer, you see more and more places where your past 
>>experience
>>with Microsoft's poor design and implementation have led you to 
>>believe that
>>certain horrible parts of using a computer are somehow inevitable.  
>>Liberation
>>from those problems sometimes seems too good to be true.
>>
>>These are considered 'good practice' for desktop systems running Windoze:
>>
>>  * Defragging.
>>  * Rebooting once a day to flush out memory that Windoze doesn't free 
>>up.
>>  * Reinstalling the operating system periodically to "clean things up".
>>  * Rebooting after a major program crashes just in case it corrupted
>>    something.
>>  * Running regular virus scans.
>>  * Not opening attachments on email.
>>
>>...none of them are needed under Linux - but that's not because there is
>>special magic inside Linux - you don't have to do those things under BSD
>>or IRIX or Solaris or HPUX either.  They are all caused by inept
>>design and implementation from our buddies at M$.
>>
>>Joe Public doesn't realise that computers can be any other way - so
>>these things are never identified as "unnatural".  People think that
>>defragging is "a good thing" - when in fact they are just exercising
>>a really lame workaround for something I'd consider to be a bug.
>>
>>So, welcome to the brave new world!
>>---------------------------- Steve Baker -------------------------
>>HomeEmail: <sjbaker1 at airmail.net>    WorkEmail: <sjbaker at link.com>
>>HomePage : http://web2.airmail.net/sjbaker1
>>Projects : http://plib.sf.net    http://tuxaqfh.sf.net
>>           http://tuxkart.sf.net http://prettypoly.sf.net
>>
>>
>>_______________________________________________
>>https://ntlug.org/mailman/listinfo/discuss
>>
>>.
>>
>>    
>>
>
>
>
>
>_______________________________________________
>https://ntlug.org/mailman/listinfo/discuss
>  
>
>------------------------------------------------------------------------
>
>This e-mail and any files transmitted with it are for the sole use of the intended recipient(s) and may contain confidential and privileged information. 
>If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. Any unauthorised review, use, disclosure, dissemination, forwarding, printing or copying of this email or any action taken in reliance on this e-mail is strictly prohibited and may be unlawful.
>Visit us at http://www.cognizant.com
>  
>