Solved - HDD's having thousand's of errors...

[zenmak]

Hi All.  I bought a Supermicro X8DAH+-F and put dual Xeon X5690's in it.  With only 1 drive attached to the mobo and during initial testing, the thing screams.. and then I put the LSI controller in it...

 

In my old server I had a LSI 9211-8i card in it and 8 drives attached.  Some of my drives are old and I understand that some of them have issues and I'm ok with it as I have an active backup.  What I can't understand is that on the old server I may have had the occasional sector relocation on some drives, etc, but I never had any "errors". 

 

On the new system, within seconds of the array being started, I see thousands of errors and if I pull up a log I see "unaligned partial completion" along with I/O errors.  Something that also had me concerned was that some of the drives came up unmountable.  I thought, new system, so I would just do a New Config and erase everything and start from scratch, at this point I don't care, I have my data already backed up, so it's not that big of a deal.  I'm wondering if the LSI card isn't compatible with the Supermicro board?  Sounds odd, but I don't know.  It worked fine in my other system...

 

Things I've tried:  

 

Installed LSI board to another slot (same issues)

Tried a different old controller and that seemed to work and didn't have any errors listed.

Mobo seems to work fine, but there is only 6 ports and I need more than that, so I need an HBA.

 

My question is, is it possible that there is something wrong with the LSI card and it's errors are only just now showing up on the Supermicro board?

What card should I be looking at for a replacement.

 

Thanks!

 

Jack

[JorgeB]

Please post your diagnostics: Tools -> Diagnostics

[zenmak]

Since I posted, I've made sure that AHCI was on in the BIOS, even though that really shouldn't affect the LSI card... Also, I went back to basics and went with a blank unraid server and also fdisked all drives to start from nothing.  Even with that, with it not being formatted or anything is still throwing errors.  My guess is faulty card.  I've attached the diags as you requested though.  Thanks for looking, I appreciate it!

 

Edit:  I should also note that the one drive that is directly plugged into the Mobo has zero errors on it...  I'm willing to accept a bad card as a result, but unsure why in one server there are zero errors and put the card with drives into another server and suddenly errors.

 

tower-diagnostics-20180912-1949.zip

Edited September 13, 2018 by zenmak

[JorgeB]

I'm seeing a lot of SMBUS errors, LSI controllers can sometimes have issues with that in some boards, there should be jumpers to disable SMB to PCI slots on the board, do it and try again.

[zenmak]

Hi. Thanks for the suggestion.  They were already disabled.  Just for kicks, I tried enabling them, still same issue.

 

Edit:  Also booted up into Ubuntu and drives appeared to be fine until I ran a benchmark test, same issues appeared.  So, I know it's not an unraid issue at least.  I've ordered 1 new replacement breakout cable to see if by some chance both breakouts are bad (doubtful).

Edited September 13, 2018 by zenmak

[JorgeB]

Then it's likely a board compatibility issue.

[zenmak]

Was my thoughts too.. I was looking on ebay last night for a replacement.  I think I'll give the updated model a try.  I looked last night and there were people using this board with the controller I'm using, but like you said, maybe it's the FW of the LSI board/some type of weird compatibility issue (just my luck).  Will keep you posted.  Thanks for all your help!

[Frank1940]

Have a check on what the firmware version is on the LSI card.  I seem to remember seeing it as my server was booting up.  It was only up for a few seconds so you might have to use a camera to actually capture it. It could be that a firmware update might fix the problem. 

[JorgeB]

38 minutes ago, Frank1940 said:

Have a check on what the firmware version is on the LSI card.

That's a good point, I forget to check, you're using a firmware with known issues (p20.00.00), though doubt it you cause so many errors, still update to latest (p20.00.07) and see if it makes any difference.

[zenmak]

41 minutes ago, Frank1940 said:

Have a check on what the firmware version is on the LSI card.  I seem to remember seeing it as my server was booting up.  It was only up for a few seconds so you might have to use a camera to actually capture it. It could be that a firmware update might fix the problem. 

It's on FW 20.  Forgot to mention that.  I did some some posts that said that 20 was an issue for some and to revert back to 19...

[JorgeB]

20.00.00 and 20.00.04 have issues, the one to use is the latest, 20.00.07

[zenmak]

It's on 20.00.00.00-IT.. Also, I had bought a 9211-8i and in the BIOS for the card it's listing it as a SAS9200-8e...  I'm thinking I'm just going to replace it and be done with it.  I don't have a computer here with UEFI BIOS in it in order to reflash it anyway.

[JorgeB]

19 minutes ago, zenmak said:

It's on 20.00.00.00

 

2 hours ago, johnnie.black said:

you're using a firmware with known issues (p20.00.00), though doubt it you cause so many errors, still update to latest (p20.00.07) and see if it makes any difference.

 

[zenmak]

Solved!!!  Thank you guys so much for the help.

 

I decided to reflash it following a guide I found on youtube.  I'm an IT guy by trade, and have flashed many things over the years, and usually it's not a big deal, but for some reason I had seen on this card that UEFI bios was required and looked like A LOT of work to get it flashed.  Had I had known that it was just a simple flash utility I would have done it sooner lol. 

 

Anyway, upgraded using the 9200-8e FW for the card and it was the 20.00.07 like you suggested.  While flashing, the utility said "Mfg Page 2 Mismatch Detected" and it fixed it on it's own.  After verifying the flash, shut the system down and hooked the cables back up again.  I'm now getting NO ERROR'S!!  And, I'm seeing speeds that I've never seen before.  Guessing that card was messed up all this time.

 

Thanks again!