Dear Experts,

I woke up this morning to find my external Maxtor drive (reiserfs) with problems.
ls reported "reading directory .: Input/output error" and fsck gave lots of "bad_indirect_item: block 20807694: The Item [212 214 0x1 IND(1)] has the bad pointer (31) to the block (3760329415) etc all over the screen.

When trying to remount the disk is not even recognized as reiserfs.

It's an only 6 months old external Maxtor USB-disk of 160GB and it holds all my backup.

Is this a common problem in Linux? The disk was 74% full when this happened and has not been making any bad noices etc.. I am running Slackware 10.1 with a 2.6 kernel with no particular modifications.

Is reiserfs so unstable? Is Linux so unstable, or should I have stayed away from Maxtor harddrives?

Please help if you can.

Thank you for reading.

-Y1

ReiserFS is a very stable filesystem. It's been well tested by many people for a long time. Probably ReiserFS and Linux are not the culprits here -- though no-one is perfect.

Either your hard drive has died. That's very bad luck, considering it's only 6 monthes old, but that can happen.

Or your filesystem is corrupt, but the hard drive itself is OK. Filesystem corruption can happen if you make certain mistakes that are particularly easy to do with external drives. Did you always unmount the hard drive before unplugging it ? (and wait for the unmounting to finish ?)

Probably the hard drive died. ReiserFS is very reliable. I've only had a couple of Maxtor hard drives (years ago, internal models) and neither of them lasted very long. I'm rough on the equipment, so it wasn't totally unexpected, but I have some now ancient IBM and Seagate drives still in service... Is your drive still under warranty?

Note that, as it's a USB-connected drive, there's always the possibility that the problem comes from your mainboard's USB controller, your USB plug or USB cable.

Wow! I completely missed that! Ugh, USB HD = disaster waiting to happen... I have a hard enough time getting USB mice to work reliably, I would never want to use a USB HD with Linux!

And the big problem with USB plugs is that many of them are very easy to un-plug, often leading to accidental unplugging. USB drives should be endowed with a small controller able to cleanly unmount as many kinds of filesystems as possible (thus probably running linux), and running on a battery with enough power to last 30 seconds. Given the very high price of external drives, that would be feasible. Until then, I won't purchase a USB hard disk.

EDIT :
I guess this scheme goes better with an Ethernet plugging. So that all disk access is done by the disk controller, not by the system. The system would then view the disk as a network share, and the disk controller should probably run a Samba server for compatibility.

Thank you everyone for the feedback and input.

The USB drive in questions sits put on my server in our storage room, and it has never been unplugged or switched off since I bought it 6 months ago. However, the USB problem brought to attention looks probable, as problems with the USB controller might have caused some serious trouble.

The question is what to do now.. I ran fsck and forced it to reiserfs and tried to reconstruct the inode table, but with no success.... Thus I suspect the data are totally gone. I can however, see some of the tree structure, but that's it. Most files are not readable, but some of the latest addition are..

Is there anything else I can try to fix this, or should I reformat the drive, and start using it for file transfer purposes only, and get a built-in IDE drive for long-term storage?

Any other parameters to fsck or any other tool that can help me recover some files?

Thanks again!

I googled :
recover reiserfs

that gave the following links :
http://forums.infoprosjoint.net/showthread.php?t=5821
http://www.linuxquestions.org/questi...icle&artid=197
among many others... I hope that helps.

If you stack Maxtor drives without sufficient cooling the life expentancy will be VERY SHORT as with any harddrive. I have had 2 Maxtors blow up like this, till I moved myself to Super Lan Boy case which provides 120mm fans that will blow thru your harddrives (there is sufficient air cap between the drives). Check your SMART status on the harddrives to see if they are reporting problem, you can enable that from the BIOS (typically disabled by default).

Juha

I think I know the problem - it occured precicely when my disk filled beyond 137GB. I've read somewhere that older BIOS don't support this, and when writing beyond 137GB just wrap around and start at 0 again, completely trashing the disk. This seems very likely, as the disk makes no "just crashed noises".

My BIOS is an ASUS CUV4X ACPI BIOS Revision 1006, with serial 07/25/2000-VT694Z-CUV4X.

Can anyone tell me if this motherboard will function with so big disks?

I am *completely* new to this. My understanding was that as long as I used a USB harddrive, it would not appear in BIOS (doesn't) and thus not be affected by any BIOS limitation. However, having seen it mess some but not all files up precicely at 137GB, I'm pretty sure this is what happened.

I was just on my way out the door to buy a new, internal 300GB disk when I noticed this - good thing I didn't waste my money just yet...

The computer and harddrive is in a cooled environment with 3 external fans taking care of it, even in hot summer days, so while the suggestion is good, I doubt that's the case here...

Any input, anyone? Thanks!!

Erk !!!!
Looks like you have bumped into (a lack of) support for "Ultra ATA".
BIOS update - *NOW*

ASUS are pretty good with the current gear (I have 3 x P4P800) - go get an update, or spend some dollars on new kit.

Thanks for replying syg00,

I found the latest BIOS upgrade on ASUS support web, but how do I know if it supports Ultra-ATA? I am worried about upgrading the BIOS and then either crashing my computer, or finding out that it doesn't work anyway, after spending $2-300 on a new disk.

Do you have any experience in this field? Would it be OK if I PM you for some more input?

-Y1

Will it support the big disks ??? - who knows. Generally when a BIOS comes out, it lists what it is adding.
One that old - should, but again who knows. Can't see it doing any harm.
To test it out, I try dd'ing 150 Gig of data - you'll know if it crosses the "limit".

The installs I've done have been pretty good; occasionally a bit of screwing around, but always eventually worked. They should offer a backup on the way - never needed it myself, but worth creating.
The installers are generally DOS - if you can't create a DOS boot disk, go to www.bootdisk.com, and download an image.

There are plenty of people happy to help here on LQ - we'll all be here; even if at different times of the day.

Linux ignores geometry information that is gathered from the BIOS, so Linux will recalculate the geometry information. Linux can handle up to a terabyte and counting.

I had my fair share of ReiserFS failing. Also ReiserFS has problems with LILO. When mounting ReiserFS, make sure you add notail or else corruption. I prefer people use XFS over ReiserFS because files can retrieve from a corrupt medium easier than you think with XFS. The problem using XFS is you need at least 512 MB of memory.

I have no problems with USB, so why it people pointing a finger at it. When mounting USB storage devices, include the sync option. This will make sure data that is sent to the medium is hopefully written to the disk.

You do not need to update the firmware of the motherboard unless you want better support for memory brands and be able to handle faster processors.

Yalla, one thing to note, your BIOS will not effect USB drives, only internal ones. It is more likely that the USB Mass Storage driver is the culprit, or maybe the Maxtor firmware in the usb case (unlikely). Your MB will generally not care about any drives that are not plugged into an IDE/SATA or SCSI controller. Has anybody else had problems, or know of problems/size limits with USB Mass Storage system?

Eric