P55M-UD2 RAID5 PROBLEM - A disk is ALWAYS missing after some time.

[LaSeR]

I am experiencing major problem with P55M-UD2 and a RAID5 array consisting of 3x1T Seagate Baracuda 7200.12 HDDs on Windows 7 Pro.
At first build this RAID and use Intel® Rapid Storage Technology 9.6.0.1014.
3 HDDs are connected to P55 SATA ports labeled SATA2_0/ SATA2_1/ SATA2_2.
After about a week IRST reported HDD on local port 1 missing.
On Reboot HDD was online and rebuilding process begun and finished ok.
So I thought nothing not to mention.
But again after 3-4 days, something happened. And on reboot all disks were missing.
For troubleshooting I was forced to remove all disk from sata ports, and reboot PC after I connected HDD one by one. Every HDD was connected and rebooted and IRST was reported every HDD as normal.
By searching I found Seagate released new Firmware for that drive, so I give it a try and update all drives to the latest.
Again 3 days of good work.
The 3rd day IRST reported again a problem.
HDD on local port 1 reported missing. Although HDD was brand new, thought maybe a hardware failure. So I took another brand new one, I updated to latest firmware, I replaced the disk reported missing and rebuild. Process finished with no trouble and system was again stable for another week.
A morning IRST reported again the replaced HDD as missing, and local port 1 as unknown.
I updated  Intel® Rapid Storage Technology Drivers to v 10.1.0.1008.
I changed sata port to another sata.
I have no clue what else to do.
I always concluded to have a disk dropped.

[Dark Mantis]

Hi

Unusual problem. I would suggest trying a HDD delay in the BIOS as a dropped drive is often because it hasn't initialised properly before the bot sequence has moved on. The other problem is pretty much the same but caused by the drive  especially if it doesn't use TLER. TLER waits for the drive to be ready before proceeding.

[LaSeR]

Hi mantis,

Thanks for replying.
Just came back from the system with the above mentioned errors.
New symptom today.
IRST bios reported SMART error on HDD on local port 1 SATA2_1.
But there was nothing wrong again with the disk.
Changing port SATA2_1 to SATA2_4, caused IRST bios to recognized it again “WITHOUT A SMART ERROR” as a member disk and begun Rebuilding.

As for TLER Function, I have read so much about it, but if a disk taking long to initialize IRST bios waits for it to be sure the disk is properly initialized.
Am I wrong?

I have moved again ports to raid Disks today but I am almost sure, RAID will Rebuild perfectly and after some day’s work will drop an HDD. I just wait to see if it will drop again the same disk which sits now on SATA2_2.
It always drops the same disk until now.

If you have any  other suggestions please let me know.

[Dark Mantis]

Well if it always drops the same disk even if it is in a different location then the disk must be faulty.

[LaSeR]

If Disk was faulty, it was the easy part.
It is the second disk (already replaced).

[Dark Mantis]

Just because it is the second disk don't think that it can't be faulty too. We have had instances on the forum of more than one disk after another being faulty. I agree that it is less likely but it does happen.

[LaSeR]

And the 2 disks you mentioned reported not faulty when you tested it???
The first disk I already replaced have NO ERRORS....
No SMART erros, No bad clusters NOTHING.
IT IS A HEALTHY ONE

[Dark Mantis]

No to be fair they were definitely faulty.

[LaSeR]

Is there a known problem with Write-back cache Enabled on IRSt v10

[Dark Mantis]

Not to my knowledge but try disabling it and see if it makes it more stable.

[LaSeR]

Hi Dark Thanks for  replying

I have seen from another post, a problem for RAID-1 which keep rebuilding.
AND the guy solved it by replacing a 500W supply he uses to a 1000w.

I have a 650W PS to the system which I consider it sufficient.
What do you thing?

[Peteruk]

I have a 650W PS to the system which I consider it sufficient.
What do you thing?

Are all the HDD's connected on the same line with SATA power connectors from the PSU?

Also are you overclocking your CPU?

[Dark Mantis]

I have seen from another post, a problem for RAID-1 which keep rebuilding.
AND the guy solved it by replacing a 500W supply he uses to a 1000w.

I have a 650W PS to the system which I consider it sufficient.
What do you thing?

I would have thought that it was sufficient as long as you aren't running a high end graphics card and the PSU isn't too old.
If you can list all the hardware it is supplying I could be more accurate.

[LaSeR]

No Mantis,

System is brand new (All Components)
P55M-UD2
i5-760
2x 2Gb  DDR3-1333
3x 1T Seagate Baracuda 7200.12
1x DVD-R/W
GForce 220 1Gb (No Extra Power)
PSU Coolermaster GX-650W

NO Over clocking.

HDDs was connected to the same power line, but since the last failure which happened last Thursday,
I split them on two separate lines from PSU. (Although PSU is a single Rail I gave it a try).

System is running ok since Thursday, but with not much load to read/write for the network.
Today I will allow it again to serve files to the network and see if it holds.

As for disabling write-back cache on RAID-5, is not an option as when I did that System was running extremely SLOWWW.

[Dark Mantis]

After looking at your specs I would say that the 650W power supply is powerful enough for your system.