[Webinar] Streamline your web hosting managementRegister Today

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 332
  • Last Modified:

bad entry in fs!?

Sep 12 00:00:03 central kernel: EXT2-fs error (device 08:02): ext2_readdir:
bad entry in directory #1727795: rec_len is smaller than minimal - offset=0, inode=1727795, rec_len=4, name_len=1

What the hell is this?  I run a perfect Linux server but I always get this every once in a while.  Anyone????

I run a Slackware 3.4 with kernel 2.0.35 and a buslogic BT958/Western Digital 4.3 SCSI3 HD.  BT958 bios 5.07b.

Please!
0
scottg
Asked:
scottg
  • 6
  • 5
1 Solution
 
JYoungmanCommented:
You apparently have a corrupted directory entry.  First, fsck the filesystem.  Go to single-user mode ("telinit 1") and then fsck that filesystem ("fsck -t ext2 -f /dev/sda2").  Then find inode 1727795 ("find /the-sda2-mount-point -inum 1727795").  The output of that command will be the name of the directory which contains the corrupted entry.

Recreate the whole contents of that directory from a backup.  The problem directory entry won't be appearing in directory listings or backups itself.

0
 
scottgAuthor Commented:
Fixing it is the easy part.  I need to know how to stop it from happening in the first place.  Its appears to be random.  I really don't need a corrupt /usr directory.  Usually its been something in /usr/src which I don't care about.
0
 
JYoungmanCommented:
Since you have a stable kernel and a decent SCSI system, I'd be very much inclined to suspect your memory.  I have had FS corruption caused by bad DRAM; even involving files that never really got read.  After I made useful complaints to the supplier, they changed the SIMMS I was using and the problems went away.

0
Never miss a deadline with monday.com

The revolutionary project management tool is here!   Plan visually with a single glance and make sure your projects get done.

 
scottgAuthor Commented:
I'll try replacing the memory, I'll know soon I guess.
0
 
JYoungmanCommented:
Let me know how it turns out.

0
 
JYoungmanCommented:
Hello?
0
 
scottgAuthor Commented:
I haven't responded because I'm still waiting to see if it screws up or not.  I made some changes such as reseating the memory in  different banks, slowing the SCSI transfer rate from 40 to 20MB, downgrading my SCSI bios(apparently recommended if under Linux), and tested for HD bad sectors and scanned the memory to see if there was trouble there.  No bad sectors, and the memory seems ok - although there was a small timing error with the memory, but this might have been the cheap program because all other scans(3 different ones) indicate ok.

Even know its been error free for 5 days, that means nothing.  I've had it go for 2 1/2 months before the error surfaced again. Give me about 5 more days...
0
 
JYoungmanCommented:
I reckon the reverse.  If only one of the three programs detected an error, I'd think that the other two are less discerning :-)
0
 
scottgAuthor Commented:
It appears to be ok.  Although that doesn't mean much, I've had it go longer without a problem before.  It should be ok.

How the hell to I answer your original question? I got no option to do that on this page.
0
 
JYoungmanCommented:
Heh.  Glad it seems better.  I guess you could accept this
dummy answer: "I think that you need to exchange your RAM for some that works, Sir!".

Don't forget to come back if it doesn't work after all...
0
 
scottgAuthor Commented:
Thanks.
0

Featured Post

Never miss a deadline with monday.com

The revolutionary project management tool is here!   Plan visually with a single glance and make sure your projects get done.

  • 6
  • 5
Tackle projects and never again get stuck behind a technical roadblock.
Join Now