Buffer I/O error on /dev/sdb1 (...) async page read
Since some months ago I have been seeing the following error message every time I reboot:
Code:
Buffer I/O error on dev sdb1, logical block 11, async page read BTW, I am backing up the data contained in this partition from Debian with rsync and the system is lagging a lot. I don't remember this behavior while using rsync previously but I usually let it do its job while I do other things away from the computer. Any thoughts? |
Please post the output from "smartctl -a /dev/sdb" (run as root), and please wrap that in [CODE] ... [/CODE] tags to preserve formatting.
|
Hi rknichols, thanks for your answer. This is the output of the smartctl command:
Code:
smartctl -a /dev/sdb |
Quote:
Code:
Commands leading to the command that caused the error were: |
It looks like you have a single block of 8 bad sectors at LBA 2136. The good news is that this should be easy to fix. I first need to confirm my assumptions that:
First, look at the output from "fdisk -l /dev/sdb" and confirm that partition 1 starts at sector 2048. If true, that means that LBA 2136 is at sector offset 88 in that partition. For a filesystem with the usual 4K blocksize, that would be filesystem block 11. Verify that block 11 is assigned to inode 7. Run the following (as root) to confirm (the part you type is in blue): Code:
# debugfs /dev/sdb1 Code:
dd if=/dev/zero of=/dev/sdb1 count=8 seek=88 |
Hi guys, thanks for your help so far.
rknichols, this is the output of fdisk -l /dev/sdb: Code:
Disco /dev/sdb: 298,1 GiB, 320072933376 bytes, 625142448 sectores |
Quote:
I'm not an expert on NTFS. I did try zeroing those 8 sectors in an NTFS filesystem, and Windows chkdsk did find and repair some errors. Of course my filesystem isn't your filesystem, and I really have no way to know whether any files were permanently lost or corrupted (I did it in a snapshot of a VM). You don't have a lot to lose (those sectors can't be read), so without any advice from someone more knowledgeable, that's all I can suggest doing. |
Hi again!
Windows doesn't complain about that partition, but this is a data partition I use to share my files between Windows and Linux in my dual-boot setup (Windows is installed in another partition on /dev/sda). Yesterday I ran chkdsk /F on the troubled partition from Windows and it found no errors at all. It's Linux which complains about /dev/sdb1 when booting the system, but other than that I have no problems whatsoever with this partition (all my data can be read/saved without issues from Linux and Windows). Since this is a NTFS partition, I will do some research to see what other tools are available on Windows to diagnose and repair NTFS problems. Thanks a lot for your help! |
If you do not need to save the data currently in the partition, what I would do is zero the entire partition (dd if=/dev/zero of=/dev/sdb1 bs=1M oflag=direct), then run "smartctl -t long /dev/sdb" and examine the results of "smartctl -a /dev/sdb" to see if the drive seems healthy, then make a new NTFS filesystem there (from Windows, probably).
|
I would still consider replacing the disk in question, if possible. As I have a Western Digital Caviar SE (Device Model: WDC WD2000JS-00PDB0), which has never reported any SMART errors (I have had it for about 10 or so years now, transplanted form my first PC), so the fact yours is reporting bad sectors, is not a good sign of things that may come. I can't guarantee that it will fail, but having said that, once again suggest you look at replacing it, if you can.
|
There are many transient events, both mechanical and electrical, that can cause a few sequential sectors to appear bad. Many of those are not an indication of future problems. Right now, a single block of 8 sectors is being reported as bad, and the drive has not seen any bad sectors elsewhere. Of course this drive is pretty small (320GB) by today's standards, and might be due for an upgrade. But, at this point I haven't seen anything else that would warrant replacing it. Do keep in mind that every new drive comes with the non-zero chance of infant mortality. At 22748 hours, the one is in the middle of its life.
|
I will try the suggestion to zero out the entire partition in a few days, when I have some time to do the backups and restore my data. In any case, I do regular backups of my data, so if something goes wrong, I will still be able to recover my data.
I'll keep you posted on how this goes after I zero out and reformat the partition. Thanks for your help, guys! :hattip: Cheers |
Sorry for the late reply.
This is just to let you know that after zeroing the partition and reformatting it (in NTFS), the message I was getting disappeared, so all is good. Cheers. |
All times are GMT -5. The time now is 07:36 PM. |