• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

HDDs power cycling without end

T

tsrtg

n00b
Joined
Sep 11, 2014
Messages
32
A couple of months ago I built the following all-in-one system:

CPU: Intel Xeon E5-1650v3 3.5GHz (6-core Haswell-EP)
Motherboard: Supermicro X10SRH-CLN4F (onboard LSI3008 SAS controller in IT mode)
Additional SAS controller: IBM M1015 (IT mode)
RAM: 8x16GB HMA42GR7MFR4N-TFTD (Total 128GB)
Power Supply: Seasonic Platinum-860
Boot SSD: Samsung 850 Pro (512Gb)
HDDs: 10x6TB + 2x4TB + 7x3TB (total 19 drives)
- 5xHitachi Desktar 5K3000 (5x3TB)
- 1xHitachi Ultrastar A7K3000 (3TB)
- 1xHGST Deskstar NAS 3TB (1x3TB)
- 2xHGST Deskstar NAS 4TB (2x4TB)
- 7xHGST Deskstar NAS 6TB (7x6TB)
- 3xHGST Ultrastar He6 (3x6TB)
BD-R: Pioneer BDR-206MBK
Graphics Card: ASUS GTX 750 Ti
CPU Cooler: Noctua NH-U12DXI4
Case: Fractal Design Arc XL (+2 internal HDD cages installed ghetto-style)
Case Fans: 7xGentle Typhoon 120mm 1450/1850/2150RPM + 2xFractal Design 140mm 1000RPM
UPS: APC Smart UPS SUA1500I
OS: Windows 7 Ultimate x64

I have the following problem with the system. Sometimes one or two HDDs start producing nasty sound and power cycling each couple of seconds. At this moment the "Start/Stop Counter" in SMART is increasing every couple of seconds. This happens spontaneously and continues until I shutdown the server, do some manipulation with the cables and boot the server again. For example this hard drive made ~4350 start/stops until I came home and shut down the server (in the morning the start/stop counter was about 150, now it's 4510):



First this happened about two months ago with one of the 6Tb Deskstar NAS drives. My first idea was to replace the molex splitter used to connect the drive. This seemed to help, but after some time the drive started power cycling again together with the second Deskstar NAS drive connected to the same molex splitter.

I one again replaced the molex splitter with a new one and replaced the SAS cable used to connect those drives to the onboard LSI3008 controller. This seemed to help, but after some time another drive (A7K3000) connected to another molex splitter but the same LSI3008 controller (via the replaced cable) started power cycling.

Then I decided to get rid of molex splitters and connect those drives to a SATA power cable using SATA->2xmolex splitters. This helped, but then another two drives (He6) started power cycling. I tried connecting them to the SATA power cable as well via another SATA->2xmolex splitter. But in this configuration, Windows 7 does not boot at all! It starts to boot but crashes when it is supposed to display the login prompt. If I check the status of HDDs via the controller BIOS all seems to be okay. I was able to boot Windows only after I reconnected one of those drives back to the molex cable.

Then I had no problems for a week or so but now another HDD (this time it's Deskstar 5K3000 connected to a different molex cable and to a different controller, the IBM M1015) start power cycling.

Now I am completely lost. What can be the problem? If my Seasonic power supply not enough for this number of hard drives? But according to the UPS, power usage is very low. And my previous server worked with 16 hard drives and 800W Odin GT power supply for years without any issues. Are all of my power cables defective? Can't be.

The temperature of hard drives is okay (35-45 degrees Celcius), I monitor that constantly.

This is nasty because I believe endless power cycling is bad for the drives, and when one of the drives enters this condition I can't do anything remotely (rebooting does not help, only physical meddling with the cables help). Please help.
 
I believe is the power distribution problem

simple is buying Supermicro 4U with sas2 backplane expander :p -> http://www.ebay.com/itm/Supermicro-...sis-2x-1200Watt-PSU-and-Rail-Kit/321646195393
send "make offer" for $250 :D... + shipping.

I have one 2U PSU has power distribution problem, since only 2(+2) molex cables from PSU.
each cables handles 8 HDs, when 12 HD is installed. some HDs get I/O erros and resetting (clicking sound)..
getting tired for that, I just install/move to 3U Supermicro with sas2 backplane, everything is good now.
 
As an eBay Associate, HardForum may earn from qualifying purchases.
Since this thing happened again, it looks like Seasonic Platinum-860 is not able to handle 20x 7200RPM HDDs reliably.

Can one recommend an ATX power supply which would not require server case and would be able to handle 20x 7200RPM hard disks? No more Seasonic please.
 
PCPower&Colling

that brand is about the only modular style PSU i can actually recommend. and you will still have to use several splitters to get all 19 drives, try to not run a splitter off of another splitter.

really as suggested (cantalup), you need a server class psu rail or psu distribution block. it is a distribution issue, but the PSU i suggested and VERY careful wiring should get you running. (your not over power spec, per se.)
 
Last edited:
How about using staggered starup? The drive cant get enough 12V supply to spin up the drives but making them spin up a couple at a time, it might work. I think there is a bios option to do it.

(rebooting does not help, only physical meddling with the cables help).

Thats a lose contact for the 12V wiring.. Have you thought about soldiering them together? staggered startups will help with this problem as well. Lose contact but still begins to startup means not enough power because of it.
 
Staggered startup is already on, but the disks do not enter this condition on startup, it happens at random time during normal operation (last time I had 72 days uptime before it happened, so I almost started to think that the problem disappeared by itself).
 
You must log in or register to reply here.
Back
Top