Supermicro CSE-216 Backplane issue?

[JimA]

Hello Folks,

First post here, and am hoping y'all can help me figure this one out. I recently purchased a Supermicro CSE-216 and X8DTU-F motherboard/chassis on eBay(https://www.ebay.com/itm/Supermicro-CSE-216-24-Bay-2U-Server-Chassis-w-X8DTU-F-Dual-LGA1366-Motherboard-/372161723547?) and started loading it up with RAM and CPUs. The issue I am running into is the RAID volume I create keeps losing 3-4 drives after a couple of hours. I have 8x 5TB Segate(https://www.amazon.com/Seagate-Barracuda-2-5-Inch-Internal-ST5000LM000/dp/B01M0AADIX). The HDD failures seem random. For example on the first failure, drives 1,3,6 disappeared. The next failure was drives 2,3,5. The randomness doesn't point me in the direction that there is something wrong with the HDDs.
I've tried several RAID cards so far to create a Volume out of the 8x5Tb drives:

• Adaptec 76105, tried RAID6
• Adaptec 72405, tried RAID6
• LSI 9211-8i, tried RAID10(no RAID 6 supported by this card), This one lasted almost 12 hours until 3 drives disappeared.
• Going to try LSI 9260-16i in RAID6 next....

So to add even more complexity to this, I have a LSI 9211-8i who is handling drive bay 0-3. I am running my OS off this LSI and running ESX 6.0 Server with 4x15k 146Gb drives in RAID10. This has yet to fail.

Right now, my thoughts point to a bad backplane model, BPN-SAS-216A. But I am unsure how to prove it or rule it out.

I appreciate everyones time looking at my post. Thanks in advance.

 

[i386]

I would try different disk as these are drives that use pmr AND smr tracks* and require special adapter/os support.

*Seagate Introduces BarraCuda 2.5” HDDs with Up to 5 TB Capacity

 

[JimA]

Do you know of any 4-5Tb 2.5" SATA drives?

 

[nthu9280]

Have you tried a different SAS cables? I don't know how to test the SAS cables but I stick to Intel/HP/Supermicro/Dell OEM cables hoping their quality standards are better than no name cheap cables.



Sent from my Nexus 6 using Tapatalk

 

[cheezehead]

Also suspecting the drives here but to rule-out the slots since you have 4x10k drives that are stable in the first four slots just cycle those through the rest of the slots to rule it out. While doing that also switch the SAS cable each time to verify the cabling as well.

As those are shingled drives, I'm not surprised. Others have had good luck with the older generation 4TB drives (https://forums.servethehome.com/ind...kup-plus-4tb-drive-cheap-2-5-4tb-drives.9539/) which were PMR.

ST4000LM016 - PMR
ST4000LM024 - SMR (4TB version of what you have)

Outside of traditional raid arrays they work fine. Since you are likely stuck with the drives at this point, i'd take a look at using ZFS and re-flashing one of the controllers over to HBA mode. It's a bit more forgiving in regarding SMR.