I have 9 Data Drives, 1 parity, 2x2 ssd cache pools

I recently upgraded my mainboard and CPU to an intel i5 9600kf and a Gigabyte B350m DS3H. I also added a gtx 1060 6gb for transcoding, though i haven't set that up yet. The HBA is now in the Pcie x4 slot, which should be ok, as far as I know. I don't run fancy stuff, so I don't care about reduced maximum bandwidth.

Since the change I get these CRC errors across multiple different disks.

In the beginning I thought it might be a power issue, as I only have a 500W PSU and the GPU might use too much 12V juice, but even that seemed unlikely, as the GPU was barely used (only a monitor for BIOS). So I took another PSU (300W, old) shorted it with a paperclip and added 6-7 drives to it. It was stable for a few days, even managed to do a parity rebuild (don't ask). Then, during another parity check, which I did to make sure everything was ok, it got through 80% when a drive dropped out, as it could no longer be read. It was still detected, but Unraid disabled it due to CRC errors which caused problems.

So my next step was that maybe something's wrong with the Molex to Sata adapters that I am using, as I heard they often are of bad quality. At this point I had to jump a second, old, 300W PSU to run all the drives (lol). Still no luck.

I guess the next step would be to change the HBA 1xminiSAS to 4xSATA connectors, but I read that they are surprisingly durable, despite their look and feel. Or maybe the problem is the HBA? I don't think it's the drives, as other SMART errors are stable and not increasing (some have pending/reallocated sectors, but they haven't increased in a long time, so it should be ok, right?).

drive 9 is the one that was disabled

The drives that are throwing errors right now are:

drive 2 - has the sata cable that used to be on drive 9 when it failed. I wanted to see if it was the cable. Maybe it was, or maybe it's the HBA. It's a 4 way splitter and the other 3 are ok

cache - is connected directly to the onboard sata

cache 2 - is connected to the HBA, but from the other 4 way splitter. Again, the other 3 are fine.

As I still have a disabled drive and the array is unprotected right now, I wanted to first ask here. Parity rebuild seems impossible, as long as the drives keep spitting out CRC errors, if only out of fear that another drive ends up disabled. How should I proceed?

I appended two syslogs and diagnostics. One of the logs is from before I did some troubleshooting.

frattininas-syslog-20251003-2300.zip frattininas-syslog-20251003-1512.zip frattininas-diagnostics-20251004-0104.zip

Edited October 3, 2025Oct 3 by Fedeöä

The first thing I would do is to replace the miniSAS cable, since CRC errors are most often the result of a bad cable.

3 hours ago, JorgeB said:

The first thing I would do is to replace the miniSAS cable, since CRC errors are most often the result of a bad cable.

That was my thought as well, but now the errors are on both miniSAS cables AND even on normal sata cables, one drive each.

Note that CRC errors don't reset. Are you seeing new ones?

16 minutes ago, JorgeB said:

Note that CRC errors don't reset. Are you seeing new ones?

Yes I know and yes they increase.

I have now unplugged and reconnected ALL Sata and ALL power from ALL drives, mixing the cables around to organize them and try to find the broken ones.

Three things happened:

1. Drives that are connected to the mainboard cannot be connected to the HBA, as that changes their names and unraid can't handle that. It thinks that the, are new drives. The other way around as well.

2. The parity doesn't even get detected by the HBA, no matter what I try, so it has to remain connected to the mainboard.

3. Now the ones throwing CRC errors are different, which seems to imply the cables are broken. I'm trying to find out which ones exactly it is.

Starts to sound like the HBA itself could be faulty. Rare that so many drives/cables would fail at the same time, but the common point connecting all of them...

What is it, when you changed mobo could you maybe have changed things so it doesn't have proper cooling anymore?

Edited October 4, 2025Oct 4 by Kilrah

1 hour ago, Fedeöä said:

as that changes their names and unraid can't handle that

That suggests you are using a RAID controller, not a true HBA.

Recommend getting one from this list; they won't have that issue.

https://forums.unraid.net/topic/102010-recommended-controllers-for-unraid/

27 minutes ago, Kilrah said:

Starts to sound like the HBA itself could be faulty. Rare that so many drives/cables would fail at the same time, but the common point connecting all of them...

What is it, when you changed mobo could you maybe have changed things so it doesn't have proper cooling anymore?

I hope not. I'm not sure it is, as one of them is not connected to it, but directly to the mainboard.

It has a heatsink, no fan, like before. Maybe a little less airflow because of more cables, but the case and HDD locations are the same. Also, problems pop up now, while the case has no sidepanels on, and when the system has only been on for minutes.

18 minutes ago, JorgeB said:

That suggests you are using a RAID controller, not a true HBA.

Recommend getting one from this list; they won't have that issue.

https://forums.unraid.net/topic/102010-recommended-controllers-for-unraid/

I am using a Broadcom / LSI MegaRAID SAS 2008 [Falcon] (rev 03) card in IT mode. It has been working for a few years without problems until now. I can see both SMART and Temps no problem. It's also working for the other 6 connected drives.

I ordered new cables and will see when they get here.

EDIT: I confirmed that one end of the 4x splitter cable is at least the source of one of the problems. When changing it from one drive to the next, the CRC errors follow it.

Edited October 4, 2025Oct 4 by Fedeöä

1 hour ago, Fedeöä said:

It has a heatsink, no fan, like before.

These do need some airflow, they're meant to be actively cooled by the airflow a server chassis provides. It's pretty common to have heat-related issues, people often 3D print a mount and install a small fan.

Edited October 4, 2025Oct 4 by Kilrah

4 hours ago, Fedeöä said:

It has a heatsink, no fan, like before. Maybe a little less airflow because of more cables, but the case and HDD locations are the same. Also, problems pop up now, while the case has no sidepanels on, and when the system has only been on for minutes.

Side panels off can hinder airflow if its a quality case (the better cases are actually designed with airflow in mind).
I printed a fan mount and added a fan to my HBA just so I wouldn't ever have to worry about it.

Edited October 4, 2025Oct 4 by whipdancer

13 hours ago, Fedeöä said:

card in IT mode

The controller is suing the Megaraid driver,. meaning it's not in IT mode, hence the issue:

02:00.0 RAID bus controller [0104]: Broadcom / LSI MegaRAID SAS 2008 [Falcon] [1000:0073] (rev 03)
	Subsystem: Dell PERC H310 [1028:1f78]
	Kernel driver in use: megaraid_sas
	Kernel modules: megaraid_sas

IT mode uses the mpt3sas driver.

16 hours ago, JorgeB said:

The controller is suing the Megaraid driver,. meaning it's not in IT mode, hence the issue:

02:00.0 RAID bus controller [0104]: Broadcom / LSI MegaRAID SAS 2008 [Falcon] [1000:0073] (rev 03)
	Subsystem: Dell PERC H310 [1028:1f78]
	Kernel driver in use: megaraid_sas
	Kernel modules: megaraid_sas

IT mode uses the mpt3sas driver.

Oh, wow. I bought it a few years ago and the seller told me it was in IT mode. It even had the tape on the pins, so I never bothered to check. It's almost weird that I never had any problems and that all drives showed up and worked flawlessly in Unraid.

Not seeing the real drive identifiers should have been an instant red flag

Just to close the topic:

I replaced one of the miniSAS->4x SATA cables. I also replaced 2 of the SSD's. They were samsung T5 portable drives that were hooked up to SATA by removing them from the case and using an adapter. They seemd to work fine for months but suddendly wouldn't stop throwing errors no matter what.

It now seems to work just fine.

I am not replacing the HBA. Even if it's not how it's "supposed" to be, it works fine and I don't want to buy another one.

Edited October 20, 2025Oct 20 by Fedeöä

I too battled through this stage with my NAS, only traced down to a specific slot that causing the issue, ended up it was a backplane issue rather than storage controller or SATA cables. Once the backplane replaced the counter stay static as is and not increasing.