Large-partition external USB disk leads to errors after usage

[joneall]

I could not think of a good title for this.

What happens is, I have an external USB disk I use for backup. I put a 300GB ext4 at the beginning and an 100GB NTFS partition farther on. I use rsync for backups. This worked fine for a couple days, then every time I tried to mount the disk, I see many, many occurences of this in /var/log/kern.log

Jul 5 18:29:15 jon-desktop kernel: [40064.658572] sd 5:0:0:0: [sdd] Unhandled sense code
Jul 5 18:29:15 jon-desktop kernel: [40064.658578] sd 5:0:0:0: [sdd]
Jul 5 18:29:15 jon-desktop kernel: [40064.658581] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
Jul 5 18:29:15 jon-desktop kernel: [40064.658584] sd 5:0:0:0: [sdd]
Jul 5 18:29:15 jon-desktop kernel: [40064.658586] Sense Key : Medium Error [current]
Jul 5 18:29:15 jon-desktop kernel: [40064.658590] sd 5:0:0:0: [sdd]
Jul 5 18:29:15 jon-desktop kernel: [40064.658594] Add. Sense: Unrecovered read error
Jul 5 18:29:15 jon-desktop kernel: [40064.658597] sd 5:0:0:0: [sdd] CDB:
Jul 5 18:29:15 jon-desktop kernel: [40064.658599] Read(10): 28 00 00 03 28 78 00 00 08 00
Jul 5 18:29:15 jon-desktop kernel: [40064.658610] end_request: critical target error, dev sdd, sector 206968
Jul 5 18:29:15 jon-desktop kernel: [40064.696323] sd 5:0:0:0: [sdd] Unhandled sense code
Jul 5 18:29:15 jon-desktop kernel: [40064.696329] sd 5:0:0:0: [sdd]
Jul 5 18:29:15 jon-desktop kernel: [40064.696332] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
Jul 5 18:29:15 jon-desktop kernel: [40064.696335] sd 5:0:0:0: [sdd]
Jul 5 18:29:15 jon-desktop kernel: [40064.696337] Sense Key : Medium Error [current]
Jul 5 18:29:15 jon-desktop kernel: [40064.696342] sd 5:0:0:0: [sdd]
Jul 5 18:29:15 jon-desktop kernel: [40064.696344] Add. Sense: Unrecovered read error
Jul 5 18:29:15 jon-desktop kernel: [40064.696348] sd 5:0:0:0: [sdd] CDB:
Jul 5 18:29:15 jon-desktop kernel: [40064.696350] Read(10): 28 00 00 00 08 78 00 00 08 00
Jul 5 18:29:15 jon-desktop kernel: [40064.696360] end_request: critical target error, dev sdd, sector 2168

I took the disk to my local store who tested it and said the disk is kaput. So I had it replaced with a 1TB disk and partitioned it similarly, but with the ext4 at 500GB. Same result, it works for a while, then dies.

I used Gparted to destroy the partition and create a new one in ext2, which I have used with success on other disks, but with smaller file systems (<200GB). Same result. Btw, the NTFS partition is not affected by this.

I can't believe this can be due to the controller in the box containing the disk. Is there some problem with rsync with large partitions?

[oshunluvr]

I've never had any issues with rsync and partition sizes. What's the enclosure you're using? Have you tried ntfs on both partitions or just extX? Are the errors exactly the same as with the old drive?

[luckyone]

I wouldn't recommend using anything but Ext2 on an external USB drive. I'm with oshunluvr on using rsync. I regularly use it through Clonezilla to backup entire HDDs and Partitions.

Have you tried running e2fsck on the Ext partition? It sounds like it might be something leftover from the Ext4 partition.

[joneall]

Does rsync work well with an NTFS partition?

[jlittle]

Does rsync work well with an NTFS partition?

It works, but you lose some metadata, like ownership.

Regards, John Little

[Teunis]

Hmm, since about 6 months I have a 1.5 GB 2.5" USB3 HD with 4 partitions, the largest is formatted to ext3, two others to ext4 and one swap.
It's set up to be bootable from the smaller partitions, the big one (1.2 TiB) is used for a data back-up with among others a 500GB Truecrypt container.
The brand is CnMemory and so far it has not given any problems.

What worries me about your report is that ntfs does not seem to be affected.