LSI 9207-8i and UDMA CRC Error Counts


Recommended Posts

Hello! I got a little problem and I don't know how to troubleshoot it. I wanted to free up some SATA ports on my motherboard and so I purchased the LSI 9207-8i SAS card, as it was on the recommended hardware list. Now, I'm getting UDMA CRC Error counts popping up on every drive.

My array includes 2 parity and 5 data disks, and has been rock stable for almost a year until I added this SAS card. Is there a setting that needs to be modified, or is this a sign of a faulty controller?

 

I'm running the follwing:

i5-7500 on a Asus Prime B250M-C motherboard

32Gb RAM

Unraid 6.8.0-rc6

 

Thanks in advance.

Link to comment
7 minutes ago, trurl said:

These are typically caused by bad connections.

 

Go to Tools - Diagnostics and attach the complete diagnostics zip file to your NEXT post.

I did purchase a pair of SAS to 4 SATA cables off of Amazon. Could both cables be faulty? I'm not just seeing errors on half the drives, but on all of them.

r0de0nas-diagnostics-20191126-0306.zip

Edited by Scheeringa1
Added info
Link to comment

I've got a new 9207-8i that I put in to an unRIAD box tonight, and same problem..... 5 out fo 5 drives on it are throwing up 199 UDMA CRC error count.  The other 3 drives using the motherbaord ports, are all fine.

 

Both of the SAS to 4x SATA cables cables I'm using are brand new and from our IT wholesaler.  The LSI card, from Ebay.  Wonder if we do have some fake cards

  • Like 1
Link to comment

OK, to anyone else who is having issues, here is where I'm at so far with troubleshooting.

I figured my first step would be to try accessing the card's BIOS and upgrading the firmware. Hard resets saw an impossibly long (2+ min) screen which read

"Avago Tehnologies MPT SAS2 BIOS
MPT BIOS-7.39.00.00"

 

I put the card in my daughter's computer, and I didn't get any splash screen. It was viewable in the motherboard BIOS (Asus Gryphon) but the text wasn't aligned properly and therefor unreadable. Windows 10 found the card and installed driver's for it without incident.

 

Put the card back into my server, and after the impossibly long boot screen I finally got the ability to view it in my server's BIOS (Asus Prime B250M-C) and it said it has controller SAS2308_2, firmware 20 IT mode.

 

Hopefully this info helps others. I've been trying to run the DOS utility in FreeDOS, but I keep getting a PAL error. And I can't manage to get EFI shell to work because of some comparability module.

 

Link to comment
15 hours ago, Davin S said:

I've got a new 9207-8i that I put in to an unRIAD box tonight, and same problem..... 5 out fo 5 drives on it are throwing up 199 UDMA CRC error count.  The other 3 drives using the motherbaord ports, are all fine.

 

Both of the SAS to 4x SATA cables cables I'm using are brand new and from our IT wholesaler.  The LSI card, from Ebay.  Wonder if we do have some fake cards

It definitely sounds like we're dealing with the same issue. Please let me know if you find a solution.

Link to comment

I have the same issue also. I have 3 different makes of cable and all 3 are giving the same error on both ports. I'm using WD 10Tb shucked drives with the 3rd pin mod. Could that have anything to do with the errors or is it likely the card?

 

I got it from eBay, so I guess there's a chance it's a fake.

Link to comment
  • 2 weeks later...
On 11/27/2019 at 11:41 AM, Scheeringa1 said:

It definitely sounds like we're dealing with the same issue. Please let me know if you find a solution.

I had the ebayer send me a new card that arrived today.  Same problem, as soon as I start a pre-clear, errros rising every second.  More importantly, I threw the LSI card in a windows machine and zeroed and entire drive to work it hard.........

 

Not one single new UDMA CRC error!  In my case at least, it appears tied to unRAID, not the card.  What's the best plan of attack to prove/disprove and maybe get a fix if it proves to be in unRAID.  If I get some time, I might try throwing the card and a drive into a spare Linux Mint box I have on the bench to see if is a broader linux issue.

Link to comment
On 12/3/2019 at 10:27 AM, Scooshie said:

I have the same issue also. I have 3 different makes of cable and all 3 are giving the same error on both ports. I'm using WD 10Tb shucked drives with the 3rd pin mod. Could that have anything to do with the errors or is it likely the card?

 

I got it from eBay, so I guess there's a chance it's a fake.

Using this site (you may want to translate to get the full info), it appears my cards are both genuine.  https://www.chinahao.com/product/37654310537/

 

That's not to say they aren't factory seconds being offloaded out the back door to ebay sellers, who in turn pass them to us.

 

Anyway, as I mentioned in my above post, the errors don't seem to happen under windows, so fake or not, that may not be the root cause of the errors.

 

 

Link to comment
On 11/27/2019 at 11:41 AM, Scheeringa1 said:

It definitely sounds like we're dealing with the same issue. Please let me know if you find a solution.

OK, I migth have some answers for you.

 

I've tested the LSI 9207 on 4 motherboards, and I've noticed that:

 - both of my original ebay card and its replacement cause CRC errors when they're in 8x PCIe slots in a couple of slightly older (4 to 5 yesr old) Intel based motherboards (graphics cards were in the 16x slots). 

 - same motherboards, but zeroing every sector under Win 10, 1 board causes CRC errors, the other doesn't ?!?!?!

 - 2 newer Ryzen based systems, B450 and X570 boards, no errors under any OS.  The B450 is running a 3200G with onboard Vega graphics, so I put the LSI card in the 16x PCIe slot.  The X570 with 2700x needs a discrete graphics card, but has multiple 16x PCIe slots, so the LSI card went in the second of those and seemed happy.

 

So despite the LSI9207 being an 8x card, I can't seem to get it work in anything less than a 16x slot...... which makes no sense at all.  The other weird anomaly in my testing, is the fact that one system behaved differently when running Linux vs Windows. 

 

Sorry, so much conflicting evidence, but there seems to be a hardware compability issue, sometimes worse under different OS's, but possibly more tied to age of chipset and possibly the amount of PCIe lanes.  More conclusive evidence would come from me jamming an old PCI graphics card in on the older Intel boards, freeing up the 16x PCIe slot to try the LSI card in, but I'm about to get a divorce if I don't finish work soon. 

 

Hope that helps get you guys on the right track, and at least give you something to try.

 

 

Link to comment
18 minutes ago, johnnie.black said:

Mine are working fine on x8 slots, my boards don't have any x16 slots.

As I said, it makes no sense.  It could be combination of 8x slot with an old chipset, or more likely, just the chipset on its own.  The oldest board I tried it on was a Gigabyte GA-Z68.  What vintage/chipset si your board?

Link to comment

Hmmm, well I would have thought the X9 based board might have been a problem if any of yours were going to do it, but maybe it's something more random again.  I'm out of boards and time to test with.  I guess I've fixed mine by changing to a newer chipset, but would have liked to have given the other guys more clrity on the source of the problem.

Link to comment

Just to add , I've just got a replacement card too and still have the same issues.

 

I have 3 12Tb WD Elements drives and I successfully pre-cleared them via USB with no CRC errors. As soon as I add them to the array and start using them, the CRC error count goes up. The CRC error count increases for all my disks.

I've even seen it increase on one of my SSD Cache drives which is connected directly to the SATA port on the motherboard, although not to the same extent as the array drives.

 

I've got an Asus ROG STRIX X570 GAMING E AM4 Motherboard with an AMD 3700X

 

The card I've just swapped to is an LSI 9207-8i  although if I view this card in windows it tells me it's a 9208 8i Mustang?

 

I'm not sure what information I can give you to help resolve the problem but I've attached my diagnostics.

 

 

unraid-diagnostics-20191211-1051.zip

Link to comment
4 hours ago, johnnie.black said:

LSISAS2308: FWVersion(20.00.00.00)

This firmware has known issues, upgrade to latest (20.00.07.00)

Thanks Johnnie, I finally managed to upgrade the BIOS using this guide: https://forums.serverbuilds.net/t/updating-your-lsi-sas-controller-with-a-uefi-motherboard/131 and some of your other posts about where to get the firmware.

 

Currently doing a parity rebuild and so far, no CRC errors. Hopefully that's everything solved and I can move on to other things.

 

I appreciate the time you take to respond to posts, it really helps those of us new to unraid.

Link to comment

So here's a final update from me. I ordered a new cable and still had issues. So I returned the card and got a new LSI SAS 9211-8i from a different vendor on Amazon and not a single problem, and it was $20 cheaper. I don't know if my 9207 was a fake, or just defective, but it was definitely a bad card.

My MicroATX mobo had one PCI slot, 2 x1 PCI-x, and 1 x16 PCI-x.....so needless to say I was using the x16 slot the entire time. Also, I'm running a i5 7600 processor. Hopefully this info helps someone!

Link to comment

For anyone stumbling across this thread and trying to update a 9207-8i, it's worth noting that of the 7 firmware downloads listed for the various OSs on the Broadcom site, only 1 has the newer 20.00.07.00 firmware bundled in... 9207-8i_Package_P20_IR_IT_Firmware_BIOS_for_MSDOS_Windows

 

I tried flashing the firmware under Linux, Windows, DOS, and UEFI shell on both a new AMD B450 system, and a 5th Gen i7, and all failed.  Wasn't until I got a crusty old Core2 Duo system left over from Noah's Ark check in system, and bam, worked first time.   I only persisted because I've read others had trouble on some boards, and having to try 3 times is a bit unlucky, but don't be surprised if it doesn't work on the first system you try.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.