DNS-323 Failing to read two 4TB Drives?

[Justin Mathews]

I had 2x2TB working drives in my DNS-323. I've just swapped both of them out and replaced both with 2x4TB drives. I was expecting Alt-F to see both drives and let me format them as new. Instead, it only sees the left drive. I've tried various combinations - one drive in the left bay, one drive in the right bay. What I've found is that either 4TB drive is recognized when installed in the left bay (implying that it's not a DOA drive), the old 2TB drives are both recognizable in the right bay, but I can't get the right bay to recognize a 4TB drive. Has anyone else seen this? I've tried with both the stock D-Link 1.10 firmware and with Alt-F.

[João Cardoso]

On Monday, November 25, 2013 12:54:49 AM UTC, Justin Mathews wrote:

I had 2x2TB working drives in my DNS-323. I've just swapped both of them out and replaced both with 2x4TB drives. I was expecting Alt-F to see both drives and let me format them as new. Instead, it only sees the left drive. I've tried various combinations - one drive in the left bay, one drive in the right bay. What I've found is that either 4TB drive is recognized when installed in the left bay (implying that it's not a DOA drive), the old 2TB drives are both recognizable in the right bay, but I can't get the right bay to recognize a 4TB drive. Has anyone else seen this? I've tried with both the stock D-Link 1.10 firmware and with Alt-F.

It is clearly a hardware problem in the box, not (directly) related with disks or firmware.

The right bay integrated circuits must be working near some thresholds, and somehow that specific 4TB Brand/Model disks puts additional load on them and the threshold is exceeded.

This might also occurs for other Brand/Model disks, and has (directly) nothing to do with disk capacity. "Directly" here means that probably some integrated circuits used in the disk puts a big load on the SATA interface; if the same ICs are used on another Brand/Model disk, the same will occur.

What you can try to do it to use 4TB disks from other brand...

just my 2c

[Justin Mathews]

I allowed the left bay to format each 4TB drive individually (with the right bay empty). Once both drives had been formatted, I put a now formatted 4TB drive back into the right bay, so that both bays were populated. This time both drives were recognized when it booted.

Could it be that an unformatted drive triggers some error (or exceeds some threshold) in the right bay but a formatted drive does not? It's been running properly for half a day now, with reboots and drive i/o both being tested today.

The drives are WD Red 4TB.

[João Cardoso]

On Monday, November 25, 2013 9:45:11 PM UTC, Justin Mathews wrote:

I allowed the left bay to format each 4TB drive individually (with the right bay empty). Once both drives had been formatted, I put a now formatted 4TB drive back into the right bay, so that both bays were populated. This time both drives were recognized when it booted.

Clever trick! And I was obviously wrong.

There has been a similar report, not with 4TB drives, if I remember correctly, but with new blank disk.

Pitty that you don't have a kernel and hotplug log  when the disk was not recognized, as that could help.

Disks are detected by the linux kernel, which generates an hotplug event that Alt-F uses, so the question is to know if the drive was detected by the kernel or not and if the event was generated or not.

Although its name is "hotplug event", it is generated also at boot time.

I guess that you didn't try to hot-plug the disks in the right bay with power on? It is possible, and I do it often.

If the issue was related by a long disk spin-up time (big disks are slower to start), could that "confuse" the kernel?

Some disks, specially rack-mountable SCSI disks, have a jumper to delay boot at power-up, so that several disks powering-up simultaneously would not put too much stress on the power supply.

But no, that does not explains why the disk becomes recognizable after it was formated, so the issue has to be related with new, blank, (big capacity?) disks.

Thanks for reporting back the cure.

Could it be that an unformatted drive triggers some error (or exceeds some threshold) in the right bay but a formatted drive does not? It's been running properly for half a day now, with reboots and drive i/o both being tested today.

The drives are WD Red 4TB.

WD Red drives are more power-hungry, I believe?