Bug#625922: RE: Bug#625922: Failures with ST2000DL003-9VT166
On 15/06/11 22:15, Paul Faure wrote: I upgraded my raid disks to 4 ST32000644NS (Seagate Constellation 2TB) and I haven't had an issue since. I have also moved the cheaper ST2000DL003-9VT166 disks to a Windows XP box (in a non raid environment) and haven't seen a problem in days now. There are plenty of references online now popping up saying that green disks are not designed or supported in a raid environment. Weather or not that's because of a physical issue, or a software issue, im not sure. Paul Just a guess: One of the features of green drivers is that they park the heads every few seconds [1] without disk activity. Can this be the root of the problem? Perhaps the HDDs are slow responding when they have the heads parked (which tends to happen too much often than with normal drives) and this causes this issue. You can disable this by forcing a one-hour timeout for parking the heads with: hdparm -S 242 /dev/sdX This needs to be done at boot time each time. Perhaps an init.d script would help. You can check the SMART attribute 193 of your drives, which tells you how many times the drive has parked the head in its life. You will see that this number is far greater on such green drivers than on normal drives. smartctl -a /dev/sdX | grep 193 Regards! [1] http://forums.anandtech.com/showthread.php?t=2085685 signature.asc Description: OpenPGP digital signature
Bug#625922: Failures with ST2000DL003-9VT166
In my case the head-parking and the power-on (real power on) values are the same, seems that I have that feature disabled. 4 Start_Stop_Count0x0032 100 100 020Old_age Always - 76 9 Power_On_Hours 0x0032 086 086 000Old_age Always - 12291 12 Power_Cycle_Count 0x0032 100 100 020Old_age Always - 80 193 Load_Cycle_Count0x0032 100 100 000Old_age Always - 80 On 16/09/2012, at 00:13, Carlos Alberto Lopez Perez clo...@igalia.com wrote: On 15/06/11 22:15, Paul Faure wrote: I upgraded my raid disks to 4 ST32000644NS (Seagate Constellation 2TB) and I haven't had an issue since. I have also moved the cheaper ST2000DL003-9VT166 disks to a Windows XP box (in a non raid environment) and haven't seen a problem in days now. There are plenty of references online now popping up saying that green disks are not designed or supported in a raid environment. Weather or not that's because of a physical issue, or a software issue, im not sure. Paul Just a guess: One of the features of green drivers is that they park the heads every few seconds [1] without disk activity. Can this be the root of the problem? Perhaps the HDDs are slow responding when they have the heads parked (which tends to happen too much often than with normal drives) and this causes this issue. You can disable this by forcing a one-hour timeout for parking the heads with: hdparm -S 242 /dev/sdX This needs to be done at boot time each time. Perhaps an init.d script would help. You can check the SMART attribute 193 of your drives, which tells you how many times the drive has parked the head in its life. You will see that this number is far greater on such green drivers than on normal drives. smartctl -a /dev/sdX | grep 193 Regards! [1] http://forums.anandtech.com/showthread.php?t=2085685 -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/bc431625-d488-4229-bdfa-a7da9e108...@claunia.com
Bug#625922: Failures with ST2000DL003-9VT166
I found the following blog post that contains some useful tips about this issue: http://paul.sullivan.za.org/kernel-disables-sata-drive-under-heavy-load-action-0x6-frozen/ signature.asc Description: OpenPGP digital signature
Bug#625922: Failures with ST2000DL003-9VT166
For me it the bug has appeared on other drives (not only Seagate), with no system load at all, and without the kernel killing tasks On 16/09/2012, at 01:53, Carlos Alberto Lopez Perez clo...@igalia.com wrote: I found the following blog post that contains some useful tips about this issue: http://paul.sullivan.za.org/kernel-disables-sata-drive-under-heavy-load-action-0x6-frozen/ -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/cbe90e97-65c5-47ee-8f0c-30e09b618...@claunia.com
Bug#625922: Failures with ST2000DL003-9VT166
I upgraded my raid disks to 4 ST32000644NS (Seagate Constellation 2TB) and I haven't had an issue since. I have also moved the cheaper ST2000DL003-9VT166 disks to a Windows XP box (in a non raid environment) and haven't seen a problem in days now. There are plenty of references online now popping up saying that green disks are not designed or supported in a raid environment. Weather or not that's because of a physical issue, or a software issue, im not sure. Paul -- To unsubscribe, send mail to 625922-unsubscr...@bugs.debian.org. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ccb64ea3944145ca83ba021824648...@51woodhill.faure.ca
Bug#625922: Failures with ST2000DL003-9VT166
I'm seeing the same sort of failure with a pair of these same drives: [ 3.570733] ata3.00: ATA-8: ST2000DL003-9VT166, CC32, max UDMA/133 wired into a rather old (2007ish) motherboard: http://www.intel.com/p/pt_BR/support/highlights/dsktpboards/dg31pr Base Board Information Manufacturer: Intel Corporation Product Name: DG31PR Version: AAD97573-203 Serial Number: BTPR741000Y3 BIOS Information Vendor: Intel Corp. Version: PRG3110H.86A.0071.2010.0318.1704 Release Date: 03/18/2010 I've upgraded the BIOS in the motherboard and it didn't fix the issue (nor were any SATA interop fixes listed in the BIOS ChangeLog). I had also replaced the power supply and both cables before that. I encountered these problems running an amd64 2.6.35-28 Ubuntu kernel (the current Maverick kernel). I just upgraded yesterday to the Natty kernel, which is 2.6.38-8 to see if it addressed the problem, and I haven't seen it occur since, but given that this bug report reports it's happening in Debian's 2.6.38-8, I'm not very hopeful. I think this is probably something broken in either the drives or in the drive-to-motherboard interaction. In looking at the motherboards which are appearing here: Gigabyte P35-DS3R Supermicro X7SPA-H All these boards are specified as supporting SATA 3.0Bps, but at least mine and the P35-DS3R are pretty old; I'm not sure about the Supermicro. So I am thinking there may just be a motherboard interoperability issue. If this kernel upgrade doesn't address the problem, then I'm considering replacing one drive at a time, and then the motherboard. (I would switch drive vendor but it won't be easy to get a refund, so I'm willing to give this a bit more effort.) -- Christian Robottom Reis | [+55 16] 3376 0125 | http://launchpad.net/~kiko Canonical Ltd.| [+55 16] 9112 6430 | http://async.com.br/~kiko -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110614162944.ga18...@anthem.async.com.br
Bug#625922: Failures with ST2000DL003-9VT166
Other potentially related issues: http://www.readynas.com/forum/viewtopic.php?f=65t=51496p=306494 http://forum.qnap.com/viewtopic.php?f=182t=39893start=30 http://forums.seagate.com/t5/Barracuda-XT-Barracuda-Barracuda/ST2000DL003-Barracuda-Green-not-detected-at-BIOS/td-p/87154/page/7 -- Christian Robottom Reis | [+55 16] 3376 0125 | http://launchpad.net/~kiko Canonical Ltd.| [+55 16] 9112 6430 | http://async.com.br/~kiko -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110614180433.ga1...@anthem.async.com.br