On 12/12/05, Bedros Hanounik <[EMAIL PROTECTED]> wrote:
> I had the exact same problem on a 300GB hard drive over USB with ReiserFS

USB isn't the same as other hard drive interfaces... maybe there's
something not acting nice.  IIRC, USB took longer than ATA/IDE and
SCSI to get reasonable bus speeds, and nowadays venders are still
pretty cheap when it comes to USB performance (maybe enough power to
use a mouse and keyboard, but rarely enough for any serious work on
anything larger than a 512 USB stick)

> by running resiserfsck, the tools suggested running it again with
> --rebuild-tree option and it was a mistake.
>
> reiserfsck --rebuild-tree screwed things  up big time; but I had a backup
> copy of my stuff, so I reformatted the HD.

IIRC, --rebuild-tree is known to bring back things from the dead. 
When you reformatted the HD, did you zero it out first?  If not, then
any future rebuild-tree might screw things up again.  Just something
to consider.

> I would probably never use the option --rebuild-tree, only --fix-fixable.

Yeah, probably a good idea... although if your reiserfs is in a state
where it's telling you to --rebuild-tree, it's time to dump it to
another disk or to CD-RWs or DVD or something and reformat, zero out,
and restore.

> I'm not sure if the partition actually was corrupted. I was able to access
> it, but on start when mounting the partition it gave me error messages;
> something like "file system not clean"

IIRC this is an internal thing - it doesn't mean there are errors on
it per se, it just means that you didn't shut down properly at least
once and the machine didn't mark the fileystem as clean - there may or
may not be an incomplete write => error on your disk.  So long as you
don't get any major errors, you should be okay.

>
> On 12/10/05, Thomas Raschbacher <[EMAIL PROTECTED]> wrote:
> > Hi,
> >
> > Finally got my new HDD (300 GB UDMA133 :D no usb anymore ..)
> > i backed up my data and ran reiserfsck on one of the other partitions
> > which were supposedly damaged.
> > it didn't work out (stopped at same point always)
> >
> > then i made a backup copy of my backup copy and ran reiserfsck on that and
> > it worked just fine.
> >
> > -> it seems that my filesystems never had errors in the first place but
> > running reiserfsck on my USB hdd caused the problems in the first place.
> > fortunately i could recover all my digital camera photos (which i will
> > definitely back up on DVD from now on!
> > some mp3's lost the filenames but who really cares about that anyway.
> >
> > i suggest adding to the reiserfsck docs that running it on usb-hdds might
> > cause problems due to usb hdd driver bugs or whatever .. or at least a
> > note that if it fails on usb-hdd copy it to a non usb hdd and try again.
> >
> > Regards,
> >   Thomas R
> > P.S.: I never enforced bandwidth allocation. but i still dont' know what
> > the problem is with this stupid USB2 storage stuffs .. it worked fine with
> > some old kernel (either early 2.6 or still 2.4 ..) any ideas what can
> > cause this kind of problems with usb hdds? or should i post to some kernel
> > mailing list/newsgroup?
> >
> >
> > --
> > -----BEGIN GEEK CODE BLOCK-----
> > GCS/CC/E/M/MU/S d- s: a--- C++++(++) UL++++ P+ L++++ E W+++ N+++ o-- K w--
> O
> > M-- V- PS+ PE-- Y++ PGP+++ t+++ 5+ X- R tv b++++ DI- D+ G++ e-->+++++ h--
> !r
> > z-
> > ------END GEEK CODE BLOCK------
> >
> > On Thu, October 20, 2005 19:55, michael chang said:
> > > On 10/19/05, Thomas Raschbacher <[EMAIL PROTECTED]> wrote:
> > >> Actually I did send it SIGUSR1 signals to get the progress so it really
> > >> seems to work fine.
> > >> I intend to get another HDD (internal) and see if i can copy the FS and
> > >> then fix it there.
> > >> how would I best copy the filesystem? DD ?
> > >> (The PRoblems I had before with this disc are USB-storage related where
> > >> the disc does stop working if one reads/writes too much data at once on
> > >> USB2)
> > >
> > > The USB2 data loss thing sounds like a bandwidth-allocation problem;
> > > since it's not always advisible to enforce it by default (well, I
> > > don't know about that, but I don't enforce it in any of my hand-built
> > > kernels).  What you are doing next sounds reasonable, please inform us
> > > how well it goes.
> > >
> > > --
> > > ~Mike
> > >  - Just my two cents
> > >  - No man is an island, and no man is unable.
> > >
> >
> >
> >
>
>


--
~Mike
 - Just the crazy copy cat.

Reply via email to