buffer I/O error logical block kernel "2.6.31.6"

damgar · 11-20-2009, 11:23 PM

I compiled a new kernel for slackware 13 yesterday and all is going well, except for the fact that during boot I get errors regarding a second hard drive. They are of the form

PHP Code:


			
buffer I/O error logical block SOMENUMBER sdb

I wasn't getting this error with the previous 2.6.29.6 kernel. As far as I can tell everything is working fine, but the errors make me nervous as this is where all my data is stored. The partition was previously formatted ntfs-3g and I thought maybe I'd done something with ntfs support while configuring the kernel so I backed up the data, reformatted to ext4 which matches all other linux partitions except for swap of course, and I get the exact same errors.

2 questions here:

1. Is there something different in this kernel in reporting errors or maybe something I set in configuration?

2. Can anyone recommend a good (simple and ubiquitous) cli program to check the disk for errors so that I can run it in both the old and new slack kernels as well as mandriva on the same machine?

Thanks

damgar · 11-21-2009, 12:14 AM

I booted into single user with each kernel and ran

PHP Code:


			
fsck -f -t ext4 /dev/sdb1

(that's a whole disk partition) and there were no errors, just that the drive is 40% non contiguous. However when I run

PHP Code:


			
fsck -f -t ext4 /dev/sdb

I am told that the superblock is probably corrupt, but I think that has to do with improper use of the command. I am definitely leaning towards operator error whild configuring the new kernel, but If anyone understands more what is going on here I'd REALLY like to hear from you.

Thanks.

XavierP · 11-22-2009, 12:34 PM

As requested, moved to Linux-Kernel

damgar · 11-22-2009, 12:44 PM

After some googling and experimenting, I've looked into /var/log/syslog and found that there are no errors to this effect when booting into any other kernel. I'm stumped since as posted earlier, everything works fine. Below are the actual lines from /bar/log/syslog.

PHP Code:


			
Nov 22 10:23:08 dtest kernel: end_request: I/O error, dev sdb, sector 240121516
Nov 22 10:23:08 dtest kernel: Buffer I/O error on device sdb, logical block 30015189
Nov 22 10:23:08 dtest kernel: Buffer I/O error on device sdb, logical block 30015190
Nov 22 10:23:08 dtest kernel: Buffer I/O error on device sdb, logical block 30015191
Nov 22 10:23:08 dtest kernel: Buffer I/O error on device sdb, logical block 30015192
Nov 22 10:23:08 dtest kernel: Buffer I/O error on device sdb, logical block 30015193
Nov 22 10:23:08 dtest kernel: Buffer I/O error on device sdb, logical block 30015194

I'm still not sure if this is a real problem or self-inflicted annoyance?

voyciz · 11-23-2009, 08:12 AM

I once had a somewhat similar problem that was cured with a low level format. Although it may be impractical, maybe it'll work.

damgar · 11-23-2009, 08:04 PM

Never having used my own kernel previous to this, is it maybe that I enabled a more verbose boot, or more likely that I screwed something up on the hardware side causing the kernel a headache when intializing the disk?
It just seems odd that with 2 distros and multiple kernel options for them, only the one I made throws the errors.