[SOLVED] Unable to detect fifth SATA disk on M4A89TD PRO/USB3

TheOneKEA · 09-06-2017, 07:52 PM

I recently reconfigured my Linux desktop to add a pair of WD Red 2.0TB hard disks to serve as a large storage array. I used an external eSATA enclosure with two bays in it to build a RAID1 array, and then installed the disks inside my desktop and connected them to the motherboard, an ASUS M4A89TD PRO/USB3.

When I booted my system with kernel 4.11.4, I observed the following:

Code:

[    1.825508] ata1: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe100 irq 19
[    1.825510] ata2: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe180 irq 19
[    1.825512] ata3: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe200 irq 19
[    1.825514] ata4: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe280 irq 19
[    1.836472] ata5: SATA max UDMA/133 abar m8192@0xfe8fe000 port 0xfe8fe100 irq 34
[    1.836473] ata6: DUMMY
...
[    2.147646] ata5: SATA link down (SStatus 0 SControl 300)
...
[    2.285785] ata4: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
...
[    2.301853] ata4.00: ATA-10: WDC WD2002FFSX-68PF8N0, 81.00A81, max UDMA/133
[    2.301856] ata4.00: 3907029168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
[    2.316865] ata4.00: configured for UDMA/133
[    2.317046] scsi 3:0:0:0: Direct-Access     ATA      WDC WD2002FFSX-6 0A81 PQ: 0 ANSI: 5
[    2.317403] sd 3:0:0:0: Attached scsi generic sg3 type 0
...

Obviously, this makes the md driver very unhappy:

Code:

[    4.710061] device-mapper: raid: Failed to read superblock of device at position 1
[    4.711989] device-mapper: table: 254:6: raid: Failed to run raid array

I shut down and verified that the power and SATA connections were secure, then booted and checked in the BIOS to verify that the disk was detected and that the BIOS was aware of it. I then booted my system again and got the same result.

I've checked the specs and the AMD SB850 southbridge on this motherboard has six native SATA III ports, and the fifth port appears to be securely seated and undamaged. Power appears stable according to lm-sensors so I'm not certain that the PSU is flaking out and causing the disk to fail to be detected by the kernel.

Does anyone have any ideas?

EDIT:

I checked my BIOS more closely and realized that the SB850 SATA controller was configured to use "IDE" mode, emulating a PATA controller. Since PATA controllers are only supposed to offer four connections for PATA drives, the BIOS decided to expose the remaining SATA ports but configure the controller to leave their links turned off.

With the controller now in AHCI mode for all six SATA ports, this is what I see in dmesg:

Code:

[    1.809897] ahci 0000:00:11.0: AHCI 0001.0200 32 slots 6 ports 6 Gbps 0x3f impl SATA mode
[    1.811897] ata1: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe100 irq 19
[    1.811899] ata2: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe180 irq 19
[    1.811902] ata3: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe200 irq 19
[    1.811904] ata4: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe280 irq 19
[    1.811906] ata5: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe300 irq 19
[    1.811908] ata6: SATA max UDMA/133 abar m1024@0xfe3fe000 port 0xfe3fe380 irq 19
...
[    2.116417] ata6: SATA link down (SStatus 0 SControl 300)
...
[    2.270305] ata4: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[    2.270354] ata5: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
...
[    2.283171] ata4.00: ATA-10: WDC WD2002FFSX-68PF8N0, 81.00A81, max UDMA/133
[    2.283181] ata4.00: 3907029168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
[    2.284146] ata5.00: ATA-10: WDC WD2002FFSX-68PF8N0, 81.00A81, max UDMA/133
[    2.284148] ata5.00: 3907029168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
...
[    2.298311] ata4.00: configured for UDMA/133
[    2.299324] ata5.00: configured for UDMA/133
...
[    2.304379] scsi 3:0:0:0: Direct-Access     ATA      WDC WD2002FFSX-6 0A81 PQ: 0 ANSI: 5
[    2.305015] scsi 4:0:0:0: Direct-Access     ATA      WDC WD2002FFSX-6 0A81 PQ: 0 ANSI: 5
...

The device-mapper subsystem is still cranky for some reason, but it still allowed the md driver to start the array:

Code:

[    4.126720] device-mapper: raid: Loading target version 1.10.1
[    4.136217] device-mapper: table: 254:4: raid: Failed to run raid array
[    4.154753] device-mapper: table: 254:4: raid: Failed to run raid array
[    4.177154] device-mapper: table: 254:4: raid: Failed to run raid array
[    4.193466] device-mapper: table: 254:10: raid: Failed to run raid array
[    4.205044] device-mapper: table: 254:10: raid: Failed to run raid array
...
[    6.648316] md/raid1:mdX: active with 2 out of 2 mirrors
...

I'm going to upgrade to kernel version 4.13 to ensure I'm running the latest versions of libata, device-mapper and md, but so far everything appears to be working correctly. I'll be interested to see if there's a performance improvement for any of my disks.