On 5/22/05, Leonard Chatagnier <[EMAIL PROTECTED]> wrote: > Adam Fabian wrote: > > >On 5/20/05, Leonard Chatagnier <[EMAIL PROTECTED]> wrote: > > > > > > > >>more than 5 months old and has never been defragged but I'm still very > >>much a newbie. > >> > >> > > > >It's very unlikely that your filesystem needs defragmenting after 5 > >months. As other have already noted, ext3 isn't prone to > >fragmentation. I wasn't even aware there was an ext2 defragmenting > >utility, though. The traditional solution would be to use > >dump/restore. > > > > > > > Ahhhh! That sounds like something that will read the hard drive and > write back each file to the drive in a > contiguous manner removing lost drive space something akin to MS > defrag. Am I correct in my assumption? > After scanning man dumpe2fs, I see mention of tune2fs and still wonder > if I have to remove some ext3 fs features > in order to run dump2fs on an ext3 fs. Do I? There were a couple more > dump programs, dump2 and dump, I believe were the names. From the man, > I believe you mean the dump2fs as it mentions ext3 in it. Am I correct > here? > If use of dump/restore is safe besides being the traditional solution, > I'd be grad to try it out. Is it relatively safe for > a newbie to use? Are there options I should use to assure not screwing > up my HD? > Thanks for your response and for seeing my request to copy my email > address. It makes my response much easier to do. > Leonard Chatagnier
Well, I trust *BSD's dump/restore completely. (I know more about *BSD than Linux.) When I say "traditional," I mean that it's what UNIX system administrators did on those extremely rare occasions when fragmentation became a problem. I know a guy that used to do it on to big reels of tape. Under Linux, I might be inclined to use GNU tar, mke2fs, and untar the the files back onto the filesystem. It's a waste of time with 4.5% fragmentation, though. I don't even think Windows utilities recommend you defragment with less than 10% fragmentation, as a rule. (Though the filesystems are dissimilar enough that I don't necessarily mean to imply that I know a comparison of this sort is relevant.) -- Adam Fabian ([EMAIL PROTECTED])
