test SATA hard disk for errors

computersavvy · 11-23-2020, 09:17 AM

Every error displayed has the same sector count and address. To me that indicates a physical defect in the drive.

With that said, however, those errors were all recorded just 4 operational hours before the current hours and no sectors show as relocated so it may be a momentary glitch or it may be the start of a major failure. I would suggest watching it, but keeping current backups just in case.

rknichols · 11-23-2020, 09:55 AM

What is the filesystem on partition 3? The Bad Block HOWTO shows, for some filesystems, examples of identifying affected files and rewriting bad sectors to cause them to be reallocated.

It's really unfortunate that this drive does not report the Current_Pending_Sector count. That count would indicate whether it is worthwhile to go through the fairly lengthy steps in the Bad Block HOWTO for each bad sector.

kosmo333 · 11-23-2020, 02:50 PM

sdb3 ext4 filesystem.

rknichols · 11-23-2020, 05:41 PM

Quote:

Originally Posted by kosmo333

sdb3 ext4 filesystem.

The HOWTO instructions for ext2/3 should still apply.

Sector offset = 241896648-7010304 = 234886344
FS block = 234886344/8 = 29360793
Running "debugfs /dev/sdb3" and then its "testb 29360793" command should reveal which, if any, inode is using the block. You can then use the "icheck" command to reveal which named file is affected.

Note that if the bad sector is actually holding filesystem metadata, "testb" will indicate no inode is using the block (it only checks for mapped data blocks). Determining just what metadata is using that block is difficult, but you can just zero the bad sector and let fsck figure out and correct what is now wrong with the filesystem.