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

RE: Lots of lost+found files ...



> On Thu, Dec 06, 2001 at 12:11:06PM -0000, Andy Bianchi wrote:
>  
> > On investigation this directory appeared to be empty but when I tried
> > to copy the missing files back in I got lots of I/O errors.
> 
> What sort of IO errors?  If you've got genuine disk IO errors then
> they will show up in the logs.

Sorry, should have been more specific - NOT actual messages in the 
log, errors from file utils, i.e. cp claimed there were errors 
trying to write to the directory.

There were no 'scarey' messages about disk I/O failing and sectors 
going ga-ga or anything like that.  The errors were limited only 
to trying to write files into the directory that was broken.

In fact, now I come to think about it, I did something along the lines
of "touch fred" in one of the directories and this created a file okay,
trying to remove one of the files that wasn't there, or actually create
it using cp caused problems.

I did run debugfs and I could still see directory enteries for the files
with weird properties, e.g. huge filesize, bizarre inode numbers.
 
> > I had a trawl through the ext3-users archive and found some mention
> > in a message from Stelian Pop, Subject: Re: 
> ext3-2.4-0.9.4-246ac5 problem...
> > which seems like the same symptoms.
> 
> So far it looks like there is a small but significant number of
> laptops which just don't have reliable disk systems.  In particular we
> know for sure of errata against certain disks which implies that they
> simply don't flush their cache to oxide on powerdown or suspend.  No
> amount of journaling will protect you if the disk is discarding data!

As far as I am aware the laptop hasn't been powered down and I don't think
I've had it suspend either ... hmmm, well certainly not while it was
doing the compile since the machine was running a screen saver when I
came back to it with the display still on!

I suppose it is possible that it could have been suspended at some point
over the past month or so.  If this was the case what sort of state could
the filesystems be in?  Would there be inconsitencies between the journal
and directory structure? 

I have to admit I don't fully understand what the journal contains and
the process used for updating the filesystem so I'm a little unsure as
to how, if at all, ext3 could create the symptoms I'm seeing.

>From what I understand though it doesn't look like whatever zapped my
disc did it in a random way like a hardware failure or something, and,
as I said the Win2k partition seems okay and I've not noticed any 
problems (Win2k thrashes around much more than Linux does anyway so I
assume it'll hit the disk even more ;-) 

> What laptop and disk is it you're using?

It's a Sony VIAO PGC-Z600HEK, with 128Mb memory and 20Gb HDD - 10Gb for
Win2k and 10Gb for Linux.

I do have an aged desktop machine at home which has vanilla RedHat 7.2
also with ext3 on a 40Gb HDD, and that has, for the past three days,
being running lots of kernel compiles (needed to check my RAM was 
okay) and I've not actually seen any problems there, maybe there is
or has been a glitch with my laptop?  

Cheers,

Andy





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