Re: failed command FLUSH CACHE EXT (was: Re: via 8237 sata errors)

[Thomas Fjellstrom <tfjellstrom@strangesoft.net>]

On June 2, 2010, Robert Hancock wrote:
> On Wed, Jun 2, 2010 at 3:10 PM, Thomas Fjellstrom
> 
> <tfjellstrom@strangesoft.net> wrote:
> > Ok, more testing, I've moved the drives over to the p35 machine semi-
> > permanently, and after a day or so of uptime I got some new errors:
> > 
> > ata3.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
> > ata3.00: failed command: FLUSH CACHE EXT
> > ata3.00: cmd ea/00:00:00:00:00/00:00:00:00:00/a0 tag 0
> >         res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
> > ata3.00: status: { DRDY }
> > ata3: hard resetting link
> > ata3: link is slow to respond, please be patient (ready=0)
> > ata3: SRST failed (errno=-16)
> > ata3: hard resetting link
> > ata3: link is slow to respond, please be patient (ready=0)
> > ata3: SRST failed (errno=-16)
> > ata3: hard resetting link
> > ata3: link is slow to respond, please be patient (ready=0)
> > ata3: SRST failed (errno=-16)
> > ata3: limiting SATA link speed to 1.5 Gbps
> > ata3: hard resetting link
> > ata3: SRST failed (errno=-16)
> > ata3: reset failed, giving up
> > ata3.00: disabled
> > ata3.00: device reported invalid CHS sector 0
> > ata3: EH complete
> > end_request: I/O error, dev sdc, sector 0
> > sd 2:0:0:0: [sdc] Unhandled error code
> > sd 2:0:0:0: [sdc] Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
> > sd 2:0:0:0: [sdc] CDB: Write(10): 2a 00 00 00 07 a7 00 00 08 00
> > end_request: I/O error, dev sdc, sector 1959
> > Buffer I/O error on device dm-0, logical block 189
> > lost page write due to I/O error on dm-0
> > end_request: I/O error, dev sdc, sector 0
> > end_request: I/O error, dev sdc, sector 0
> > JBD2 unexpected failure: do_get_write_access:
> > buffer_uptodate(jh2bh(jh)); Possible IO failure.
> > 
> > end_request: I/O error, dev sdc, sector 0
> > end_request: I/O error, dev sdc, sector 0
> > ------------[ cut here ]------------
> > WARNING: at /home/damentz/src/zen/main/linux-
> > liquorix-2.6-2.6.34/debian/build/source_amd64_none/fs/buffer.c:1199
> > mark_buffer_dirty+0x74/0x90()
> > Hardware name: P5K SE
> > Modules linked in: nfs lockd fscache nfs_acl auth_rpcgss sunrpc ipv6
> > acpi_cpufreq cpufreq_ondemand freq_table cpufreq_conservative
> > cpufreq_userspace cpufreq_powersave af_packet ext3 jbd loop
> > snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_pcm_oss
> > snd_mixer_oss snd_pcm rtc_cmos rtc_core snd_timer tpm_tis nvidia(P) tpm
> > rtc_lib tpm_bios evdev snd intel_agp pcspkr asus_atk0110 soundcore
> > i2c_i801 snd_page_alloc button i2c_core processor dm_mod raid10
> > raid456 async_raid6_recov async_pq raid6_pq async_xor xor async_memcpy
> > async_tx raid1 raid0 multipath linear md_mod ext4 mbcache jbd2 crc16
> > usbhid sd_mod ata_generic pata_acpi uhci_hcd ata_piix libata floppy
> > scsi_mod thermal atl1 mii ehci_hcd [last unloaded: scsi_wait_scan]
> > Pid: 3283, comm: jbd2/dm-0-8 Tainted: P          
> > 2.6.34-0.dmz.8-liquorix- amd64 #1
> > Call Trace:
> >  [<ffffffff8103bf23>] ? warn_slowpath_common+0x73/0xb0
> >  [<ffffffff81101cd4>] ? mark_buffer_dirty+0x74/0x90
> >  [<ffffffffa005a3c9>] ? __jbd2_journal_unfile_buffer+0x9/0x20 [jbd2]
> >  [<ffffffffa005d8a3>] ? jbd2_journal_commit_transaction+0xba3/0x12d0
> > [jbd2] [<ffffffff810542d0>] ? autoremove_wake_function+0x0/0x30
> >  [<ffffffffa0061701>] ? kjournald2+0xb1/0x210 [jbd2]
> >  [<ffffffff810542d0>] ? autoremove_wake_function+0x0/0x30
> >  [<ffffffffa0061650>] ? kjournald2+0x0/0x210 [jbd2]
> >  [<ffffffff81053e3e>] ? kthread+0x8e/0xa0
> >  [<ffffffff81033e8d>] ? schedule_tail+0x4d/0xf0
> >  [<ffffffff81003c94>] ? kernel_thread_helper+0x4/0x10
> >  [<ffffffff81053db0>] ? kthread+0x0/0xa0
> >  [<ffffffff81003c90>] ? kernel_thread_helper+0x0/0x10
> > ---[ end trace c90e4c710c9ef513 ]---
> > end_request: I/O error, dev sdc, sector 0
> > 
> > (and plenty more dmesg lines from lvm and ext4/jbd2 screaming about the
> > io commands failing)
> > 
> > I take it that this means the drive is likely pooched? I'm going to try
> > some more tests, and make sure both of the WD drives are on their own
> > power cable first. but I'm betting now that the drive is just failing.
> > This would make 2 out of 4 in the same batch that had issues. The
> > first one would increase the sector reallocated count 4 every hour or
> > so. Now this one fails a flush cache command (and other spurious
> > errors).
> > 
> > I guess its time to break out the WD diagnostics disk.
> 
> I think it's a fairly safe assumption there's something wrong with the
> drive - it looks like the drive just pretty much stopped talking..

I've only managed to see that error once though. The last few times I've 
booted that machine I get the same old DMA error messages I posted before.

Unfortunately I haven't been able to run the WD Diagnostics thing on it so 
far, it either takes an age to load, or hangs at the screen prior to the 
license screen.

I seem to have rather bad luck with hard drives. Every time I buy more than 
two, I tend to get one or two failures out of the batch. 25-50% failure rate 
almost. Horrible. I at least average 1 dead hard drive a year since I got my 
first computer.

-- 
Thomas Fjellstrom
tfjellstrom@strangesoft.net