Adaptec ASR-8885 Raid 5 degraded after disks power cycled by controller

[perdrix]

I've got a RAID 5 behind an Adaptec ASR-8885. Yesterday I set the card up to power the drives down after 20 minutes of inactivity.

Nov 16 16:13:59 charon [1109]: [12120] Task priorty changed from Low to High: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), logical device 0 ("Shared"), UniqueID E05CE835.
Nov 16 16:48:29 charon [1109]: [927] Logical device power management state changed to Off: .
Nov 16 16:48:29 charon [1109]: [180B blob data]
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 1, WWN: 50000C0F01C85E41, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F1054265, firmware level: VR02.
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 0, WWN: 50000C0F0129DEFD, vendor: WD, model: WD4001FYYG, S/N: WMC1F1532925, firmware level: D1R5.
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 5, WWN: 50000C0F010E7B41, vendor: IBM-ESXS, model: WD4001FYYG-23S, S/N: F0E0X2F0, firmware level: XA39.
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 6, WWN: 50000C0F0129C691, vendor: WD, model: WD4001FYYG, S/N: WMC1F1528634, firmware level: D1R5.
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 4, WWN: 50000C0F01DEB949, vendor: WD, model: WD4001FYYG, S/N: WMC1F1533218, firmware level: D1R5.
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 7, WWN: 50000C0F01FF3D59, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F0E1VFXS, firmware level: VR08.
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 3, WWN: 50000C0F023D7B31, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F0D99DRX, firmware level: VR08.
Nov 16 16:54:14 charon [1109]: [426] Power management state changed to Powered off: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 2, WWN: 50000C0F01D0D7C1, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F0918722, firmware level: VR07.
Nov 16 16:54:14 charon [1109]: [191B blob data]
Nov 16 17:09:34 charon [1109]: [927] Logical device power management state changed to Active : .
Nov 16 17:09:34 charon [1109]: [180B blob data]
Nov 16 17:10:04 charon [1109]: [426] Power management state changed to Full rpm: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 1, WWN: 50000C0F01C85E41, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F1054265, firmware level: VR02.
Nov 16 17:10:04 charon [1109]: [426] Power management state changed to Full rpm: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 0, WWN: 50000C0F0129DEFD, vendor: WD, model: WD4001FYYG, S/N: WMC1F1532925, firmware level: D1R5.
Nov 16 17:10:04 charon [1109]: [426] Power management state changed to Full rpm: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 5, WWN: 50000C0F010E7B41, vendor: IBM-ESXS, model: WD4001FYYG-23S, S/N: F0E0X2F0, firmware level: XA39.
Nov 16 17:10:04 charon [1109]: [426] Power management state changed to Full rpm: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 4, WWN: 50000C0F01DEB949, vendor: WD, model: WD4001FYYG, S/N: WMC1F1533218, firmware level: D1R5.
Nov 16 17:10:04 charon [1109]: [426] Power management state changed to Full rpm: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 7, WWN: 50000C0F01FF3D59, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F0E1VFXS, firmware level: VR08.
Nov 16 17:10:04 charon [1109]: [426] Power management state changed to Full rpm: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 3, WWN: 50000C0F023D7B31, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F0D99DRX, firmware level: VR08.
Nov 16 17:10:04 charon [1109]: [426] Power management state changed to Full rpm: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), channel: 0, deviceID: 2, WWN: 50000C0F01D0D7C1, vendor: WD, model: WD4001FYYG-01SL3, S/N: WMC1F0918722, firmware level: VR07.
Nov 16 17:10:04 charon [1109]: [191B blob data]
Nov 16 17:10:04 charon [1109]: [301] Logical device is degraded: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), logical device 0 ("Shared"), UniqueID E05CE835. Repair as soon as possible to avoid data loss.
Nov 16 17:10:09 charon [1109]: [180B blob data]
Nov 16 17:16:05 charon [1109]: [925] Logical device power management poweroff setting changed from 20 minutes to never: controller 1 ( Adaptec ASR8885 #4D39138FB5F Physical Slot: 16 ), logical device 0 ("Shared"), UniqueID E05CE835

Firmware level is 7.16.0 b33456 (latest).

Has anyone else tried to spin the drives down on these?

David

 

[kapone]

Yes. I use it all the time (on 71605, 78165 among others).

1. Don't use RAID-5, use RAID-6 (obligatory warning, but in your case probably pertinent)
2. Don't mix drives from manufactures/different firmwares in the same RAID set.
3. At least one of your disks has failed.

 

[perdrix]

1. Don't use RAID-5, use RAID-6 (obligatory warning, but in your case probably pertinent)
2. Don't mix drives from manufactures/different firmwares in the same RAID set.
3. At least one of your disks has failed.

1. Tedious but slow process to convert - can't add an extra drive as server is full
2. Strange I've heard precisely the opposite advice - is this one of those cases where different views hold?
3. Not before the test - the array was "Optimal"

 

[kapone]

1. Tedious but slow process to convert - can't add an extra drive as server is full
2. Strange I've heard precisely the opposite advice - is this one of those cases where different views hold?
3. Not before the test - the array was "Optimal"

Perfectly acceptable, but spinning a disk down might have broken it. "might" is the keyword here. How old are these disks? Do you know the load/unload counts on these?

 

[perdrix]

The load/unload count on the device in question is 78927, one drive has 86499, several of the others are under 30 cycles.