SoftRAID 8.3 having issues with MacOS Sequoia 15.1

Save a SoftRAID tech support file with the disks connected and attach it.

Certainly something is going on.

Hi SoftRAID Support,

The file is attached.  Took me two tries as the machine locked up the first try.

---KenH

@kenh 

I notice you have a couple older drives, and one of them is not reporting SMART.

I suspect that is your failing drive.

You have two 14TB and 212TB drives, the 12TB are past their expected lifetime of 25,000 to 30,000 hours.
Take a look at the drive with 38,000 hours, it is not showing SMART, which is a sign of failure (Unless you specifically told SoftRAID to ignore SMART checks on that disk)
I would make sure you are backed up, then remove that disk. See if you still get panics. Then plan on replacing both drives with new ones, the faulty one first.

I used Hard RAIDs in the past, before I switched to SoftRAID not that long ago.  I've had drive failures before, and I just get notices on which drive that needs replacing, then the system goes into protected mode.  Is this normal for a drive failure to make the OS crash for SoftRAID?  No errors or warnings?  This seems to be an incredibly bad system.

---Ken

Hi,

Not sure if my problem is the same as yours but I've been experiencing continuous panics with my ThunderBay 6 and 8 enclosures, regardless of the ThunderBay enclosure type or drive age (eg, from a RAID set using brand new certified drives with <150 hours usage and one RAID-5 set with >60,000 hours per drive). I was able to reproduce the panic 100% of the time when copying specific files. My Mac would randomly panic or hang. It was definitely not a Mac hardware issue as Apple spent 2 weeks testing my SSD, logic board and RAM (I have a 2019 Mac Pro).

Once I disconnected my ThunderBay enclosures, I have not had a single panic in 3 weeks (versus daily panics with even 1 ThunderBay attached).

It took a while but SoftRAID support confirmed that it is an issue between the SATA controllers in the ThunderBay and Sequoia. There's actually a blog posting on OWC's website if you haven't read it already. Depending on the drive type, a Mac connected to a ThunderBay will likely crash. Drive age wasn't a factor - just the specific drive model. Even if your drive has 60,000 hours, it shouldn't cause your Mac to hang randomly without some specific warning from SoftRAID.

So for now (at least in my case), I cannot use any of my ThunderBay enclosures until this is resolved (no estimate on timing according to OWC as they said they are understandably dependent upon Apple)

Paul

@lychee128 

the issue you are referring to is a bug introduced by MacOS that should be fixed in 15.2.

the SoftRAID driver passes all IO through MacOS, we have no ability to "intercept" kernel panics.

Apple has been working hard on kernel panics and anyone who gets a panic with Sequoia (except this known Mac Pro issue) can contact us and we can collect the key information required to report this formally, so that the cause can be addressed.

Note: Having >60,000 hours per drive means your drives are ancient in terms of usage. I have seen drives like this not reporting any predictive failure alerts, but causing kernel panics by hanging at times. Drives with critical data should be replaced after 25,000-30,000 hours. After that, their reliability is suspect for critical data. Use them for off shelf archives, or non critical data only.

I've taken the one drive offline and have ordered 4 new Ironwolf Pro drives.  The three remaining drives are rebalancing now.  Once I get the new drives, I'll add them, then phase out the other drives.  You might want to give a warning to users about the MacOS Sequoia issue, so others will at least know what is going on here.  Thanks.

---Ken

@kenh 

no, drives should not cause the MacOS system to crash. There are very few crashes "caused" by SoftRAID, and when we find one, we either fix it, or report it to Apple engineering to fix it.But when a drive hangs, it can still bring down MacOS. An example is a (on an old drive, or a failing drive) SMART check. It needs to report back within 2 minutes. If it does not respond to a SMART query within two minutes, there can be a kernel panic as a result. I have a drive on my desk that does this reliably, regardless of whether initialized by Disk Utility, or SoftRAID. (its a great test drive, until it ultimately fails totally!)

MacOS with Sequoia, has dramatically reduced kernel panics as a result, as it appears to be a high priority inside Apple now.

@kenh 

Regarding warning users about the 2019 Mac Pro with Sequoia, we did write a blog post and sent an email to those signing up for urgent alerts.
Its fortunate that most 2019 MacPro users are staying with older OS versions (probably because of specific tools they are using do not support Sequoia yet), so we have not had too many complaints with this.

We discovered this during the Sequoia beta period, but it was too late to have been fixed in 15.0. We were hoping for 15.1. Users have reported that it is fixed in 15.2, however.

@softraid-support 

Yes, the RAID set that I have that's over 60K hours is a backup of a backup. It's definitely non-critical but I had an unused ThunderBay 6 so I thought I would just keep using it until it dies. Unfortunately, I had to upgrade my Mac Pro to Sequoia to address get a bug fix for the 4M2 and overlooked the issue with the SATA controllers...

1) You mentioned the SATA Controller kernel pack is fixed in 15.2. Are you referring to the latest 15.2 beta?

2) Where can I sign up for urgent alerts?

Thanks, 

Please notify me if it is fixed with MacOS 15.2, when it is released.  Thanks again.

---Ken

Yes, we have had users report that 15.2 beta fixes this issue.

We now have the critical issues mailing list sign up on our main support page:

https://software.owc.com/support/softraid /"> https://software.owc.com/support/softraid/

Signed up.  Thanks again.

PS. Having an issue removing the missing disk (i.e. the bad disk).  I tried "Remove Missing Disk" multiple times, including after a clean shutdown and restart, but the "missing disk and degraded warnings still show.  Any ideas?

===Ken

@kenh 

Probably a terminology error.
Remove Missing Disk is a term that only applies to RAID 1 (Mirror) volumes. We support up to 16 disks in a mirror and users often need to remove the off site disks from the volume. That removes them from being tracked.
Any disk still connected is not "removed".
If a RAID 4/5 volume is missing a disk, you need to "Add disk". then let it rebuild.