trouble with seagate 1.5TB - 2.6.27.4

trouble with seagate 1.5TB - 2.6.27.4

[kyle]

Hi,

Please cc me if possible.

I just built a new server with some seagate 1.5TB harddisks with mirrored
/boot and root partition , The system sometime degrade the root partition
with error message as follow, this happened 4 times within 2 days.

Oct 30 22:41:20 rsync2 kernel: ata2.00: exception Emask 0x0 SAct 0x0 SErr
0x0 action 0x6 frozen
Oct 30 22:41:20 rsync2 kernel: ata2.00: cmd
ea/00:00:00:00:00/00:00:00:00:00/a0 tag 0
Oct 30 22:41:20 rsync2 kernel:         res
40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
Oct 30 22:41:20 rsync2 kernel: ata2.00: status: { DRDY }
Oct 30 22:41:20 rsync2 kernel: ata2: hard resetting link
Oct 30 22:41:21 rsync2 kernel: ata2: SATA link up 3.0 Gbps (SStatus 123
SControl 300)
Oct 30 22:41:21 rsync2 kernel: ata2.00: configured for UDMA/133
Oct 30 22:41:21 rsync2 kernel: ata2: EH complete
Oct 30 22:41:21 rsync2 kernel: sd 1:0:0:0: [sdb] 2930277168 512-byte
hardware sectors (1500302 MB)
Oct 30 22:41:21 rsync2 kernel: sd 1:0:0:0: [sdb] Write Protect is off
Oct 30 22:41:21 rsync2 kernel: sd 1:0:0:0: [sdb] Write cache: enabled, read
cache: enabled, doesn't support DPO or FUA
Oct 30 22:41:21 rsync2 kernel: end_request: I/O error, dev sdb, sector
20691533
Oct 30 22:41:21 rsync2 kernel: md: super_written gets error=-5, uptodate=0
Oct 30 22:41:21 rsync2 kernel: raid1: Disk failure on sdb2, disabling
device.
Oct 30 22:41:21 rsync2 kernel: raid1: Operation continuing on 1 devices.
Oct 30 22:41:21 rsync2 kernel: RAID1 conf printout:
Oct 30 22:41:21 rsync2 kernel: --- wd:1 rd:2
Oct 30 22:41:21 rsync2 kernel: disk 0, wo:0, o:1, dev:sda2
Oct 30 22:41:21 rsync2 kernel: disk 1, wo:1, o:0, dev:sdb2
Oct 30 22:41:21 rsync2 kernel: RAID1 conf printout:
Oct 30 22:41:21 rsync2 kernel: --- wd:1 rd:2
Oct 30 22:41:21 rsync2 kernel: disk 0, wo:0, o:1, dev:sda2
Oct 30 22:42:29 rsync2 xinetd[2278]: START: telnet pid=20246
from=192.168.2.1
Oct 30 22:43:17 rsync2 kernel: md: unbind<sdb2>
Oct 30 22:43:17 rsync2 kernel: md: export_rdev(sdb2)
Oct 30 22:43:22 rsync2 kernel: md: bind<sdb2>
Oct 30 22:43:22 rsync2 kernel: RAID1 conf printout:
Oct 30 22:43:22 rsync2 kernel: --- wd:1 rd:2
Oct 30 22:43:22 rsync2 kernel: disk 0, wo:0, o:1, dev:sda2
Oct 30 22:43:22 rsync2 kernel: disk 1, wo:1, o:1, dev:sdb2
Oct 30 22:43:22 rsync2 kernel: md: recovery of RAID array md1
Oct 30 22:43:22 rsync2 kernel: md: minimum _guaranteed_  speed: 1000
KB/sec/disk.
Oct 30 22:43:22 rsync2 kernel: md: using maximum available idle IO bandwidth
(but not more than 200000 KB/sec) for recovery.
Oct 30 22:43:22 rsync2 kernel: md: using 128k window, over a total of
10241344 blocks.
Oct 30 22:44:51 rsync2 xinetd[2278]: EXIT: telnet status=0 pid=20158
duration=251(sec)
Oct 30 22:45:50 rsync2 kernel: md: md1: recovery done.
Oct 30 22:45:50 rsync2 kernel: RAID1 conf printout:
Oct 30 22:45:50 rsync2 kernel: --- wd:2 rd:2
Oct 30 22:45:50 rsync2 kernel: disk 0, wo:0, o:1, dev:sda2
Oct 30 22:45:50 rsync2 kernel: disk 1, wo:0, o:1, dev:sdb2

Regards,
Kyle

Re: trouble with seagate 1.5TB - 2.6.27.4

[James Andrewartha]

kyle wrote:
> Please cc me if possible.
> 
> I just built a new server with some seagate 1.5TB harddisks with mirrored
> /boot and root partition , The system sometime degrade the root partition
> with error message as follow, this happened 4 times within 2 days.

It appears that model of drive is broken:
http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=2390
http://www.techreport.com/discussions.x/15837
Apparently disabling the write cache is a workaround.

-- 
James Andrewartha

Re: trouble with seagate 1.5TB - 2.6.27.4

[Mark Lord]

James Andrewartha wrote:
> kyle wrote:
>> Please cc me if possible.
>>
>> I just built a new server with some seagate 1.5TB harddisks with mirrored
>> /boot and root partition , The system sometime degrade the root partition
>> with error message as follow, this happened 4 times within 2 days.
> 
> It appears that model of drive is broken:
> http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=2390
> http://www.techreport.com/discussions.x/15837
> Apparently disabling the write cache is a workaround.
>..

Heh, I wonder if there's been some firmware spying between Seagate and Hitachi..
because my two Hitachi 750GB drives can be made to have the same problem.

Triggers quite reliably (perhaps 1/5 attempts?) by running "hddtemp"
(to read drive temperature) during heavy I/O.  Bingo.

I imagine the smartmontools would have a similar effect.

Cheers

Re: trouble with seagate 1.5TB - 2.6.27.4

[kyle]

----- Original Message ----- 
From: "Mark Lord" <liml@rtr.ca>
To: "James Andrewartha" <jamesa@daa.com.au>
Cc: "kyle" <kylewong@southa.com>; <linux-ide@vger.kernel.org>
Sent: Wednesday, November 05, 2008 1:06 PM
Subject: Re: trouble with seagate 1.5TB - 2.6.27.4


> James Andrewartha wrote:
> > kyle wrote:
> >> Please cc me if possible.
> >>
> >> I just built a new server with some seagate 1.5TB harddisks with
mirrored
> >> /boot and root partition , The system sometime degrade the root
partition
> >> with error message as follow, this happened 4 times within 2 days.
> >
> > It appears that model of drive is broken:
> >
http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=2390
> > http://www.techreport.com/discussions.x/15837
> > Apparently disabling the write cache is a workaround.
> >..
>
> Heh, I wonder if there's been some firmware spying between Seagate and
Hitachi..
> because my two Hitachi 750GB drives can be made to have the same problem.
>
> Triggers quite reliably (perhaps 1/5 attempts?) by running "hddtemp"
> (to read drive temperature) during heavy I/O.  Bingo.
>
> I imagine the smartmontools would have a similar effect.
>
> Cheers
>

Now I have to rebuild the raid everyday since every morning when I check out
the mirrored root partition, one of the drives (not always the same) fails
with the mentioned error. Strange the the 6x1.5TB raid 0 storage partition
even down because of this.

Kyle

Re: trouble with seagate 1.5TB - 2.6.27.4

[Tejun Heo]

kyle wrote:
> ----- Original Message ----- 
> From: "Mark Lord" <liml@rtr.ca>
> To: "James Andrewartha" <jamesa@daa.com.au>
> Cc: "kyle" <kylewong@southa.com>; <linux-ide@vger.kernel.org>
> Sent: Wednesday, November 05, 2008 1:06 PM
> Subject: Re: trouble with seagate 1.5TB - 2.6.27.4
> 
> 
>> James Andrewartha wrote:
>>> kyle wrote:
>>>> Please cc me if possible.
>>>>
>>>> I just built a new server with some seagate 1.5TB harddisks with
> mirrored
>>>> /boot and root partition , The system sometime degrade the root
> partition
>>>> with error message as follow, this happened 4 times within 2 days.
>>> It appears that model of drive is broken:
>>>
> http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=2390
>>> http://www.techreport.com/discussions.x/15837
>>> Apparently disabling the write cache is a workaround.
>>> ..
>> Heh, I wonder if there's been some firmware spying between Seagate and
> Hitachi..
>> because my two Hitachi 750GB drives can be made to have the same problem.
>>
>> Triggers quite reliably (perhaps 1/5 attempts?) by running "hddtemp"
>> (to read drive temperature) during heavy I/O.  Bingo.
>>
>> I imagine the smartmontools would have a similar effect.
>>
>> Cheers
>>
> 
> Now I have to rebuild the raid everyday since every morning when I check out
> the mirrored root partition, one of the drives (not always the same) fails
> with the mentioned error. Strange the the 6x1.5TB raid 0 storage partition
> even down because of this.

Yikes, I'll ask Seagate what's going on and post it on linux-ata wiki
when I know more.

Thanks.

-- 
tejun

Re: trouble with seagate 1.5TB - 2.6.27.4

[Mark Lord]

Tejun Heo wrote:
> kyle wrote:
>> ----- Original Message ----- 
>> From: "Mark Lord" <liml@rtr.ca>
>> To: "James Andrewartha" <jamesa@daa.com.au>
>> Cc: "kyle" <kylewong@southa.com>; <linux-ide@vger.kernel.org>
>> Sent: Wednesday, November 05, 2008 1:06 PM
>> Subject: Re: trouble with seagate 1.5TB - 2.6.27.4
>>
>>
>>> James Andrewartha wrote:
>>>> kyle wrote:
>>>>> Please cc me if possible.
>>>>>
>>>>> I just built a new server with some seagate 1.5TB harddisks with
>> mirrored
>>>>> /boot and root partition , The system sometime degrade the root
>> partition
>>>>> with error message as follow, this happened 4 times within 2 days.
>>>> It appears that model of drive is broken:
>>>>
>> http://forums.seagate.com/stx/board/message?board.id=ata_drives&thread.id=2390
>>>> http://www.techreport.com/discussions.x/15837
>>>> Apparently disabling the write cache is a workaround.
>>>> ..
>>> Heh, I wonder if there's been some firmware spying between Seagate and
>> Hitachi..
>>> because my two Hitachi 750GB drives can be made to have the same problem.
>>>
>>> Triggers quite reliably (perhaps 1/5 attempts?) by running "hddtemp"
>>> (to read drive temperature) during heavy I/O.  Bingo.
>>>
>>> I imagine the smartmontools would have a similar effect.
>>>
>>> Cheers
>>>
>> Now I have to rebuild the raid everyday since every morning when I check out
>> the mirrored root partition, one of the drives (not always the same) fails
>> with the mentioned error. Strange the the 6x1.5TB raid 0 storage partition
>> even down because of this.
> 
> Yikes, I'll ask Seagate what's going on and post it on linux-ata wiki
> when I know more.
..

Now slashdotted:  http://techreport.com/discussions.x/15863