Cause of I/O error with dd after 10GiB using split?

# dd if=/dev/sda3 bs=2M | split -b 3758096K - /mnt/sdc2/DELL-BACKUP/OS.img.part- dd: error reading '/dev/sda3': Input/output error 5164+1 records in 5164+1 records out 10831069184 bytes (11 GB, 10 GiB) copied, 383.179 s, 28.3 MB/s #

This is my 4th attempt in two 32-bit Ubuntu-based os. Block size was varied. File system is NTFS:

sda3 218.2 GB (primary Windows 7 partition)
sda2 14.7 GB recovery
sda1 39.2 MB system

system & recovery are successfully archived.

Originally Posted by JASlinux # dd if=/dev/sda3 bs=2M | split -b 3758096K - /mnt/sdc2/DELL-BACKUP/OS.img.part- dd: error reading '/dev/sda3': Input/output error 5164+1 records in 5164+1 records out 10831069184 bytes (11 GB, 10 GiB) copied, 383.179 s, 28.3 MB/s #

Probably a hardware problem in the disk. What messages are logged by the system when the error occurs?

Please use [CODE] ... [/CODE] and not QUOTE tags for this sort of included material. Use of QUOTE tags makes it harder to quote your material in a reply.

Originally Posted by rknichols Probably a hardware problem in the disk. What messages are logged by the system when the error occurs?

Where is the log? Your guess seems more plausible now, as I just got the error again with 7z, having tried split with & without gzip earlier. The only constant is dd, so I thought that might be the problem.

How do I get out of this with a backup?

I used CODE first then switched.

I guess it is the same again: https://www.linuxquestions.org/quest...3/#post6474729
You should know where are the system logs located, either in /var/log or use journalctl (or dmesg).
dd itself is ok, but 1. you may use it incorrectly or 2. the disk is damaged. I guess in your case the disk is not healthy, but without details hard to say more. You can also use smartctl to check your disk.

system & recovery are successfully archived.

This means nothing unless you can successfully restore your system from these backups.