On Sat, 5 Mar 2016, Scott Holder wrote: > > If you send the output of dmesg, that would be helpful. You may see a > > scsi error when you unmount filesystems, because old scsi drives don't > > support the cache commands the kernel sends. > > I'm pretty sure it's bad sectors on the drive.
If that's the case, you would see the same sector numbers again when you scan the media (dd if=/dev/sda6 of=/dev/null bs=32k), or move the drive to a different host. > Need to get an external and/or rescue ramdrive going to do a proper scan > and repair going on it. Or remember the incantations to manually mark > sectors bad on a live drive. > > [31515.050000] sd 0:0:1:0: [sda] tag#0 FAILED Result: hostbyte=DID_BAD_TARGET > driverbyte=DRIVER_OK The DID_BAD_TARGET indicates a selection or reselection timeout. Might be worth checking the cabling and termination? Or perhaps try using a MacOS utility to scan the drive media (maybe HDT Toolkit?). > ... > [31515.090000] Buffer I/O error on dev sda6, logical block 1606594, lost sync > page write > [31515.100000] JBD2: Error -5 detected when updating journal superblock for > sda6-8. I've seen errors similar to this with the ext4 module, when the block device fails a cache flush command. Old scsi disk drives don't implement that. > ... > [31515.130000] EXT4-fs (sda6): Remounting filesystem read-only > [31515.130000] EXT4-fs (sda6): previous I/O error to superblock detected But I was using the ext4 module with an ext2 filesystem (no journal) which might explain why I never saw this result. --
