Read Errors on 5 TB HDDs

[WhenKittensATK]

My current setup:
Intel Pentium G630
Gigabyte Z68MA-D2H-B3
G.Skill 32GB
Corsair CX430
2x IBM M1015
Supermicro BPN-SAS-836TQ
5x 3 TBs HDD (4x Toshiba / 1 x HGST)
6x 5 TBs HDD (3x Toshiba / 3x HGST)

This year has been a bad streak for hard drives for me. I've RMA 4x 3 TBs and I'm about to return 4x 5 TBs. Right now I'm having issues with my 5 TBs showing read errors in Ubuntu (async page read). This error originally showed up on my 3 TBs so I removed those. Then it showed up on 3x 5 TBs I replaced it with. Then it showed on the 3x 5 TBs I replaced those with. I've replaced the HBA and SAS breakout cable. I also tried different PCI Express slots. I've bypassed the backplane. I also tried using the motherboard SATA slots. All these attempts always lead to HDDs not showing up in OS and/or async page read errors. I'm just totally lost on what to try next. I'm feeling I should just dump all the hardware and start fresh. The only thing I got working are the remaining 3 TB HDDs.

 

[tunatime]

Have you tested the drives by themselves or on another MB? 11 hdds dieing in the year seem suspicious. Are they runing hot? Have you checked the power delivery to them?

 

[michalrz]

Maybe, just maybe you're hitting a power limit of your PSU. The CX430 has 20 amps available for (both?) the 5V and 3.3V rails.
I opened up some pics of 5TB HGST and the sticker says ~0.5A from 12V AND ~0.5A from 5V. 11 drives is almost 6 amps of the 20 available on your PSU.
Anyone?

 

[WhenKittensATK]

Have you tested the drives by themselves or on another MB? 11 hdds dieing in the year seem suspicious. Are they runing hot? Have you checked the power delivery to them?

The hard drives seem to load up just fine on a Windows 10 machine with a HDD dock. Only the 5 TB Toshiba pull up warnings for SMART data for Reallocated Sector Ct ~1000 on all three. I haven't actually tried to transfer any files to them yet through Windows.

 

[mnewxcv]

Maybe, just maybe you're hitting a power limit of your PSU. The CX430 has 20 amps available for (both?) the 5V and 3.3V rails.
I opened up some pics of 5TB HGST and the sticker says ~0.5A from 12V AND ~0.5A from 5V. 11 drives is almost 6 amps of the 20 available on your PSU.
Anyone?

my first thought as well.

 

[WhenKittensATK]

I'll look into getting a bigger PSU as well.

 

[michalrz]

I'll look into getting a bigger PSU as well.

Can you try transferring part of your storage pool to another computer first? That way you could rule it out.

 

[drescherjm]

Is this just the Toshiba drives or the HGST as well?

There is a recent thread about Toshiba drives having bad sectors after some usage.

Warning possible bad batch of Toshiba 6TB from Amazon and Newegg.

 

[skiddierow]

If you're using too many splitters off SATA power, that could be the source of grief. Try putting a load on all the problem drives, and see if you can hear them power cycling.

You might want to look into PCIE splitters, or a bigger PSU.

 

[WhenKittensATK]

Is this just the Toshiba drives or the HGST as well?

There is a recent thread about Toshiba drives having bad sectors after some usage.

Warning possible bad batch of Toshiba 6TB from Amazon and Newegg.

It's just the Toshiba drives that pull up SMART data as warning status. HGST are okay so far except maybe one of them.

If you're using too many splitters off SATA power, that could be the source of grief. Try putting a load on all the problem drives, and see if you can hear them power cycling.

You might want to look into PCIE splitters, or a bigger PSU.

My PSU has 3x molex connectors. I am using two 1x molex to 2x molex splitters. The backplane uses 4 molex connectors and 1 molex for fan controller. I use two SAS to 4x SATA cables.