HDD-Errors

[Padi75]

Hi all,

I have built a Server based on:

- Supermicro X11SCH-LN4F Motherboard

- Intel Core i3-8100 CPU

- 4x Samsung DDR4 16 GB UDIMM 288-PIN 2666 MHz ECC

- 4x WD HDDs WD 8TB WD80EFAX Red (Data)

- 2x 1TB SSD Samsung 970 EVO PLUS M.2 (Cache)

- Broadcom Avago SAS9300-8i HBC

- Power Supply Be Quiet! Pure Power 11 350W 80PLUS Bronze

- Chenbro-Chassis with 2x4 HDD  Mini-SAS-HD Backplanes

- Unraid Version 6.8.3

 

The SSDs are sitting directly at the Motherboard.

The 4 HDD's are connected via the Chenbro-Backplane and a new Mini-SAS-HD to Mini-SAS-HD Cable from Supermicro with the Broadcom-HBC.

 

You can find the System-Log and the Diagnostic-Files attached.

data-serv1-syslog-20201021-0812.zip

data-serv1-diagnostics-20201021-1013.zip

 

Everything worked fine and the System runs stable, but as soon as I want to write files to the Data-HDDs I get much HDD-Errors.

Often some thousands and at two or more drives simultanously:

 

 

I think the Smart-Tests are ok. (?) - See attached pictures:

 

 

 

 

 

What I've done till now: (Two Times)

- Under Tools I've made a new config with none of the drives preserved.

- Then I've them new assigned. (Yes, in the right order - I've noted their position)

- Started the array and wrote the two parity discs with new parity information.

That took about 14 hours with no read- / writ errors nor other system-errors or warnings (see syslog)

- Finaly I had a new array with "0" errors.

- Then I tested the array by copying some files from my Win-PC to a Unraid user share (via Cache-SSDs)

- In the meantime I've seen, that the two parity disks were spun down.

- The other two data-HDD's haven't spun down - I think because shortly before I accesed the user shares.

- As soon I started the Mover manually, I got the errors on the parity drives.

 

I have checked all Cable-Connections and HDD-Hotswap-Trays.

 

I want to mention, that some weeks ago I used the same HDDs, SSDs and HBC in another Testserver with an:

- Supermicro X10SRi-F Motherboard

- registered RAM 64GB

- Intel Xeon E5-2667v3

- Case and Backplane from Supermicro

Although I wrote 7 TB to the array I had never such an issue over 8 Weeks although I started with Unraid 6.8.3 and after

a view days I switched to the Beta .25, .28 and .30 of Unraid.

 

Then I built this second server with a completely new written Unraid-USB-Stick with the stable version 6.8.3.

I just putted a few old config-files like super.dat an usershares to the stick.

 

Could it have to do with the Spin Down of the Disks ?

Should I "upgrade" to a Beta-Version? (Don't want so for important Data...)

 

Can you please help me! I'm stucking...

Perhaps there is someone, who knows the answer through the Syslog...

 

Thanks and best regards

 

EDIT:

That the two parity drives are affected seems to be pure coincidence.

Yesterday Data Drive 1 and 2 were also affected.

In one case there were errors on 3 or all 4 drives simultanously.

 

Edited October 21, 2020 by Padi75

[JorgeB]

LSI firmware is very old, try again after updating to latest.

[trurl]

For future reference, the syslog and SMART reports you attached are already included in the Diagnostics so no need to attach them separately. 

[Padi75]

2 hours ago, trurl said:

For future reference, the syslog and SMART reports you attached are already included in the Diagnostics so no need to attach them separately. 

Thank you.

[Padi75]

5 hours ago, JorgeB said:

LSI firmware is very old, try again after updating to latest.

I'll try it tonight.
Then I'll build new my parity drives and test again.

[Padi75]

Hi JorgeB,

 

like you said, I updated the BIOS and Firmware of the LSI-Controller and of my Motherboard.
Thanks a lot for your superb instruction how to flash the Controller under Unraid !

 

The parity-build finished now with no errors (like before).

Parity-Discs were spun down and I've copied a file to the system without any error.

Now it seems that all works fine! Let's have a look if it's lasting, when I'll copy more files to the NAS.

 

Your advice to update was right. I thought about that too, but wouldn't belive in that,

because the controller has run in the other server with no problems.

Thanks a lot again for your hint and instruction.