Harddrive errors?

[jaboua]

---

I was looking through my systemlogs trying to find something on why DMA was disabled on my harddrives even thought the kernel was set to enable it by default, then I found this:

Jan 23 00:40:37 pentagon ide0: reset: success
Jan 23 00:40:37 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 00:40:37 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 00:40:37 pentagon ide: failed opcode was: unknown
Jan 23 00:40:37 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 00:40:37 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 00:40:37 pentagon ide: failed opcode was: unknown
Jan 23 00:40:37 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 00:40:37 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 00:40:37 pentagon ide: failed opcode was: unknown
Jan 23 00:40:37 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 00:40:37 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 00:40:37 pentagon ide: failed opcode was: unknown
Jan 23 00:40:37 pentagon ide0: reset: success
Jan 23 00:40:37 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 00:40:37 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 00:40:37 pentagon ide: failed opcode was: unknown
Jan 23 00:40:37 pentagon end_request: I/O error, dev hdb, sector 78165360

...and that's only a part of it. It's lots of those errors. Does anyone know what may be wrong? I'm starting to get a bit worried, is the hd going to die soon? Or is it just some configuration problem?

[oz]

I've never seen that one before. Is the HD making noises, or anything? Have you run any disk tests on the drive?

[jaboua]

The hd doesn't make any specific noise no - it's been making some scratching noises since I got it thou, and that was in 2001... I haven't tried any disk test no, I'm gonna google it up. I haven't noticed any data-loss from the drive either. I also think it might be something with my kernel, after trying the ubuntu livecd, mounting the windows and linux partitions on /dev/hdb, running a dd if=/dev/zero of=/mnt/lin/testingfile, turning on DMA, repeating the dd-test and checking the logs, there are no errors:

Jan 23 13:32:40 localhost gconfd (ubuntu-23477): Resolved address "xml:readonly:/usr/share/gconf/debian.defaults" to a read-only configuration source at position 6
Jan 23 13:32:40 localhost gconfd (ubuntu-23477): Resolved address "xml:readonly:/var/lib/gconf/defaults" to a read-only configuration source at position 7
Jan 23 13:33:12 localhost gconfd (ubuntu-23477): Resolved address "xml:readwrite:/home/ubuntu/.gconf" to a writable configuration source at position 0
Jan 23 13:35:39 localhost kernel: [4295032.320000] NTFS driver 2.1.22 [Flags: R/O MODULE].
Jan 23 13:35:39 localhost kernel: [4295032.430000] NTFS volume version 3.0.
Jan 23 13:35:46 localhost kernel: [4295039.247000] ReiserFS: hdb2: found reiserfs format "3.6" with standard journal
Jan 23 13:35:47 localhost kernel: [4295040.276000] ReiserFS: hdb2: using ordered data mode
Jan 23 13:35:47 localhost kernel: [4295040.307000] ReiserFS: hdb2: journal params: device hdb2, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30
Jan 23 13:35:47 localhost kernel: [4295040.311000] ReiserFS: hdb2: checking transaction log (hdb2)
Jan 23 13:35:47 localhost kernel: [4295040.355000] ReiserFS: hdb2: Using r5 hash to sort names
Jan 23 13:35:47 localhost kernel: [4295040.357000] ReiserFS: hdb2: warning: Created .reiserfs_priv on hdb2 - reserved for xattr storage.

I think I'm gonna test the disk, if nothing wrong give the kernel26 Arch-standard kernel a try and check if the errors are still there.

EDIT: While installing kernel26, I found this after some more googling:
http://www.ussg.iu.edu/hypermail/lin...09.2/0110.html

A bit old, but it may be the same issue. Now I'll go and try rebooting with new kernel...

[jaboua]

I tried smartmontools to test the drive, here's the result:

root@pentagon smartmon]# smartctl -d ata -H /dev/hdb
smartctl version 5.33 [i686-pc-linux-gnu] Copyright (C) 2002-4 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

So I guess my drive is physically OK... On some google result I saw something about changing something in BIOS, so I'll try that too. kernel26 didn't help:

Jan 23 15:17:02 pentagon ReiserFS: hdb2: found reiserfs format "3.6" with standard journal
Jan 23 15:17:02 pentagon ReiserFS: hdb2: using ordered data mode
Jan 23 15:17:02 pentagon ReiserFS: hdb2: journal params: device hdb2, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30
Jan 23 15:17:02 pentagon ReiserFS: hdb2: checking transaction log (hdb2)
Jan 23 15:17:02 pentagon ReiserFS: hdb2: Using r5 hash to sort names
Jan 23 15:17:02 pentagon Freeing unused kernel memory: 252k freed
Jan 23 15:17:02 pentagon hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jan 23 15:17:02 pentagon hdb: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=78165360, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jan 23 15:17:02 pentagon hdb: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=78165360, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jan 23 15:17:02 pentagon hdb: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=78165360, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jan 23 15:17:02 pentagon hdb: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=78165360, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon hda: DMA disabled
Jan 23 15:17:02 pentagon hdb: DMA disabled
Jan 23 15:17:02 pentagon ide0: reset: success
Jan 23 15:17:02 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 15:17:02 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 15:17:02 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 15:17:02 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jan 23 15:17:02 pentagon hdb: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=78230639, sector=78165360
Jan 23 15:17:02 pentagon ide: failed opcode was: unknown
Jan 23 15:17:02 pentagon ide0: reset: success

[jaboua]

Nope, in BIOS I tried switching from "Other" installed OS to "Win95/2000", I tried switching disk mode from "Other" to "DOS", I tried turning off LBA (which made GRUB not start) and I cleared the ESCD - still, same error. From the log it seems like it's trying to switch on DMA before mounting, but it's failing for some reason... I dunno why. I'm tempted to try a 2.6.14-kernel too later to see if that has any effect, since it seems like some others have had different problems with 2.6.15. Please write if you have any clue, got to go now have to do some homework... Later today I might try the gentoo livecd too and see if it reports any errors. I just cleared my messages.log file and executed this, so I can see if it's reported any errors after boot too about it with or without DMA:

dd if=/dev/hdb of=/dev/null && mv /var/log/messages.log /var/log/messagesnodma.log && touch /var/log/messages.log && hdparm -c1d1 /dev/hdb && dd if=/dev/hdb of=/dev/null

[jaboua]

I stopped the DD-commands after copying about 6GB... 2 GB w/DMA and 4GB w/o DMA, nothing showed up in the log... I tried compiling the latest stable 2.6.14 kernel from kernel.org, still the same old logs. Tried the gentoo livecd, worked fine there too after mounting the partitions, no errors in the log.

[jaboua]

I tried googling even more, finally I came into a thread with about the same problems as me... Exactly the same errors. I translated from spainsh to english this site with freetranslation: http://www.linux.org.ve/archivo/l-li...er/046577.html

Your hard disks do not bear UDMA, and you are trying to use UDMA al to start the machine, treats removing him UDMA al kernel and tries again.

The translator may not be perfect, but I understand it ... I think the kernel tries to enable UDMA on my discs, but they only work with normal DMA. I'll try disabling the automatic DMA-on feature in the kernel and add a hdparm-command to the boot and see if it works better... I hope it works better.

Nope... It's still the same errors. I added a hdparm line in /etc/rc.local or something like that thou so I do get DMA on my discs, but the logs are still full of harddrive errors (and it only happen at boot).